Skip to main content

Full text of "A Statistical Account of Noun-Inflection in the Veda"

See other formats


STOP 



Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 
purposes. 

Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- 
journal-content . 



JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 
contact support@jstor.org. 



ARTICLE XI. 
A STATISTICAL ACCOUNT OF 

NOUN-INFLECTION IN THE YEDA. 

By CHARLES R. LANMAN, 

ASSOCIATE-PROFESSOR FOR SANSKRIT IN THE JOHNS HOPKINS UNIVERSITY, BALTIMORE, MD. 



Presented to the Society October 25th, 1871. 



The treatment of the inflective forms of any Indo-European 
tongue requires a division of the subject into two : the verb, 
and the noun. Bach of these, again, is bipartite : the one part 
deals with the formation of word-stems; the other, with the 
suffixes of inflection. 

For the language of the Veda, three of these four chapters 
have been made the object of special systematic investigation : 
the first, or mode and tense-stems, by Delbriick, in his book 
Das Altindische Verbum (Halle, 1874), p. 83 ff.; the second, or 
verb-inflection in a narrower sense, by the same scholar, ibidem, 
pp.23-78 ; the third, by Lindner in his Altindische Nominalbildung 
(Jena, 1878); a careful working out of the fourth, or noun- 
inflection in the Veda, has become a desideratum. This I now 
attempt to supply. 

In so doing, certain limitations of the plan of the work and 
of the ground covered by it seemed advisable. I have accord- 
ingly restricted myself to the texts of the Rigveda-sanhita and 
the Atharvaveda-sanhita : for the former I have aimed to be 
complete and exhaustive ; from the latter I have excerpted, I 
believe, all the noteworthy forms, generally excluding those 
which conform entirely to the norms of the later language, but 
including all belonging to the double sets, like those in -d and 
-dni, -dsas and -as, etc. The terms u noun " and " verb " cover 
the great bulk of the Vedic vocabulary. Under " noun " I 
have included all that is subject to distinctively nominal inflec- 
tion — that is, substantives, adjectives, and participles, and even 
the nominal forms from pronominal stems. Thus, of the three 
vol. x. 45 



326 G. R. Lanman, 

equivalent forms, upare, upards, and upardsas, I have taken 
account of the second and third, but not of the first These 
true pronominal forms, along with a few numerals, make a small 
residue of so varied character that it is best treated by itself. 
The infinitives and absolutives, too, given by Delbriick (p. 
221 ff.), I have not repeated, although, on their inflectional side, 
they are nouns. 

With the history and genesis of the forms of declension I 
have concerned myself only where circumstances demanded. 
These matters lie beyond the scope of my work ; but, I may 
add, such collections of facts as the following seem to me to be a 
sine qua non for fruitful speculation upon this subject — if, 
indeed, it ever can be very fruitful. 

An extension of this work over the texts of the Brdhmanas 
and later literature would doubtless be useful. This has been 
done for the verb by Professor Avery in the preceding article ; 
but the inflective system of the Veda, when contrasted with 
that of the later language, is comparatively less multiform on 
the field of the noun than on that of the verb : for although, 
upon both, the Veda shows a rank growth of forms which die 
out later, yet the process and result have been different for each. 
Originally there may have been some slight modal or other 
distinction between crruxvas, grnu, crnudki, cpnuhi, and crudhi ; 
but the instrumental singular feminine of an z'-stem had 
absolutely no implicit difference of function or meaning, 
whether ending in -id, -yd, 4, -i, or ind. In the later Sanskrit, 
the conception of modal relations and the ability to utilize the 
ample means of nice expression offered by its system of verb- 
inflection, instead of developing to wonderful perfection (as they 
did with the Greeks), became lost; and with them also a host 
of forms ; until, for example, the style of the Hitopadeqa hobbles 
along with the aid of bungling absolutives and participles. The 
conception of the case-relations, on the other hand, naturally 
survived, and, for each, at least one form, and generally only 
one, for its expression. The inflective sj'stem of the nouns has 
become contracted, rigid, and uniform ; but not, like that of the 
verb, essentially mutilated. 

From considerations of practical importance, the work of 
collection was based upon Grassmann's excellent Index-vocabu- 
lary to the Rig- Veda ; by it alone was almost entire complete- 
ness attainable, and the necessity of separating by exegesis 
multitudes of homonymous forms (for the most part) avoidable. 
That a careful revision of the exegesis of all these ambiguous 
forms, with constant reference to the views of Roth in the St. 
Petersburg Lexicon, would have been a proper preliminary to 
my collection of materials, I cannot deny ; but I hope that the 
progress of Vedic study may be better furthered by my offering 



Noun- Inflection in the Veda. 327 

what I have at once, than by waiting to carry out such a plan. 
It is, at any rate, a real step in advance to have all that is 
doubtful and questionable conveniently grouped together. 
This done, the cases of the docket are at least prepared for the 
decision of scholars of riper j udgment. 

The material from the Atharva-Veda was collected by the 
aid of Professor Whitney's complete manuscript Index to that 
text. To him I desire to make acknowledgments for his 
kindness in lending me the volume, and for the advice and 
encouragement he has given me in my work. 

The order in which I have treated the different stems aims 
chiefly at convenience of practical reference, and is in general 
as follows : 

I. Stems whose suffixes end in vowels ; 

II. Suffixless stems ; 

III. Stems whose suffixes end in t, n, and 8. 

I. The a-stems (m. and n.), as most numerous and important, 
come first ; then the <2-stems (f .) ; the i-stems (m., n., and f.) and 
the w-stems (m., n., and f.) follow, along with the few mascu- 
lines and feminines in i and d, not radical ; next, the ow-stems 
(m., f., and n.), and finally those in o, ai, and au. 

II. Under this head I have arranged the stems of nouns, 
simple and compound, substantive and adjective, whose only 
element, or whose final element, is a root; first come the 
vocalic, and then the consonantal stems, in alphabetical order. 
Some words, however, as agreeing in declension with class II., 
and therefore better excluded from III., are treated here, 
although formed with suffixes (as -aj, -ij, -it, ut, -vat). They 
are neither numerous nor important enough, to warrant our 
setting up special classes for them. 

III. First come the stems ending in vocalic root + t, belong- 
ing declensionally to II. ; then, the participles : namely, present 
participles in -ant (or -at of reduplicating verbs) and perfect 
participles in -vans ; and, as analogous to these, the compara- 
tives in -iyahs; and finally, the stems formed by the suffixes 
-vant and mant, -van and -man, -an and -in, -as, -is, and -us. 

Lindner has treated the form and accentuation of noun-stems 
only in relation to their derivation from roots. There are also 
certain changes of stem and of accent which are common to 
whole classes of words, and are purely declensional; as, the 
shifting of the accent to the ending in the genitive plural of 
oxytone stems in i, u, r, and the distinctions of "strong," 
" middle," and "weak " forms of the same stem. These facts 
will be noticed in their proper places. (For a general discussion 
of the Stammahst'ifende Declination, see Brugman, in Curtius's 
Studien, ix.3til ff., and Kuhn's Zeitschrift, xxiv.l ff.) Besides 



828 0. R. Lanman, 

these, there are many and varied minor peculiarities, occurring 
sporadically ; as, the nasalization of a final vowel (e. g., 
camasd'n, I.s.m., which thus becomes identical with the samhiid- 
form of an accusative plural masculine before a vowel), or of a 
stem (e. g., yuj-, yunj-), the " strengthening " of a stem (e. g., pad-, 
pd'd-), or the contraction of a case- form (e. g., vedhd'm for vedha- 
sam). The comparison of such of these as are similar will be 
facilitated by an Index. A concluding section will take account 
of words which resist convenient classification ; and will give 
comparative statistics concerning the use of double-forms, and 
so much of a general summation of results as may be necessary. 

The full set of declensional forms of many words is a piece of 
patch-work made by putting together fragments of the sets of 
several stems : so for akshdn dkshi, asthdn dstM, dadhdn dddhi, 
sakthan sdkthi ; yakdn yakr't, calcdn cdhrt ; ndJcta naktdn, cirshd 
cirshdn ; cdkshas cakshan (AV. x.2.6). This " change of theme " 
is especially common between the i and ?-stems (cf. oshadhi, 
bhu'mi, rd'tri, and their forms), so that it is often doubtful to 
which of these a form is to be referred. This subject deserves 
a special investigation. (See Asboth, Die Umwandlung der 
Themen im Lateinischen, Grottingen, 1875.) 

I have enumerated in full all the forms that occur, whenever 
it seemed that any valuable practical purpose would be served 
by such an enumeration. Grrassmann's and Miiller's Indexes 
being in the hands of every student of the Veda, it is believed 
that a detail of case-forms "found in the text on every page, and 
counting by scores and hundreds, would be a useless expendi- 
ture of space. 

The names of the cases are abbreviated by their initials in 
capitals (N. A. I. D. Ab. G. L. V.) ; designations of number 
(s. d. p.) and of gender (m. f. n.), by their initials in small letters. 
Gfr. stands for Grassmann ; his Worterbuch zum Big -Veda is 
designated by Wb., his Uebersetzung, by Ueb. The Lexicon of 
Bohtlingk and Eoth is cited by the initials BR. The pada- 
text is denoted by p. 



N.s.m.] Noun- Inflection in the Veda. 329 



STEMS IN A. 

The following is a synopsis of the terminations of the 
inflectional forms : 

Singular: N.m., -as, -a'/ A.m., -am, -dm?, -a'; N.A.n., -am, 
-a'/ I.m.n., -ad, -d, -end, -ena (-enan, -dn) ; D.m.n., -dya (-dyan); 
Ab.m.n., -at, -aat? ; G.m.n., -asya (-asyan), asiaf, asyd; L.m.n., 
-e, -a'/ V.m.n., -a (-an), -d. 

Dual : N.A.V.m., -d (-dn), -an, -a; N.A.n., -e; LD.Ab.ra.n., 
-dbhydm, -dbhidm; G.L.m.n., -ayos, -os. 

Plural: N.V.m., -dsas, -ds, -aas; A.m., -dn, -dn, -dns; N.A.V.n., 
-d, -dni, -ad ?, -a ? ; I.m.n., -ebhis, -ais; D.Ab.m.n., -ebhyas, -ebhias; 
G.m.n., -dndm, -dnaam, -dm/ L.m.n., -eshu. 

There are no irregularities of accent. 

Nominative Singulab Masculine. 

The regular ending of this case is -s. It is the most frequent 
form in the RV., and occurs 10,071 times (cf. Benfey, Vedica, p. 
115). The number of stems in connection with which it is found 
is 1845 : their frequency forbids their enumeration. The form 
indras, for example, occurs 523 times; ekas, 104; eshds, 106; kds, 
111; devds, 203; mitrds, 132; yds, 1010; sds, 891; somas, 220; 
vdrunas is found 94 times ; sutds, 93 ; pundnds, 84 ; vrshabhds, 
80; pdvamdnas, 65 ; bhdc/as, 65 ; vipras, 59; ptitds, 56; su'rias, 
56 ; jdtds, 55 ; mdrtias, 54. 

' The only alleged deviation from rule in the form of the N.s.m. 
is krdnd' (Scholiasts, after Pan. vii.1.39). It is not hard to believe, 
with Bollensen (Z.D.M.G. xxii.574), that -d may represent -as 
sporadically ; and perhaps the relation of -a(s) to -o is like that of -d 
to -an in the dual. In two instances the supposed krdnds is not 
followed by a sonant, and the exegesis by no means requires a 
nominative. I have therefore referred them to the I.s.m., q.v. 

Even after the elision of the * of the ending, crasis of the 
preceding a with a following vowel sometimes takes place. Cf. 
Kuhn, Beitrage, iv.199. Numerous examples of this may be seen 
among those cited Rik Pr. ii.33,34. Thus sdh asi ukthydh, ii.13.2, 
is to be read sd'si nkthiah. So sdh end' dnikena, ii.9.6, must read 
saind'nikena; and sd utd bdndhuh, AV. ii. 1.3, sdtd bdndhuh. In 
vii.86.4, the crasis (tureydm) is indeed metrically unavoidable, 
but does not necessitate our reading it as a vocative (Gr., Wb.), 
nor yet as a verb (root tar, tur, Gr.,Ueb.). In view of the anal- 
ogous cases just cited, I prefer to follow the pada, and read turds 
iydm : ' With sins forgiven, humbly I would straightway betake 
myself to thee.' In vi.46.4, we need not resort to an emendation 
(vrshabhd va), nor yet follow the pada, vrshabhd'-iva; but may 
regard the samA^d-reading as another instance of crasis after 
elision, vrshabhd iva. The usage of the Atharvan is instructive. 
It is given by Whitney (note to Ath. Pr. ii.56). The contraction 



330 G. R. Lanman, [a-stems. 

of -as iva to -eva is here the rule (in 46 out of 59 instances). 
Compare also Benfey, Abh. d. kdn. (res. d. Wiss. zu Gottingen, xix. 
250 ff. 

The -as of the N.s.m. appears in the samhitd with exceeding 
frequency as -o (see A.Weber, Seitrdge, iii.385 ff.). In this final o 
other case-forms are sometimes concealed. Thus x.155.2, cattd is to 
be resolved into cattd' u, N.s.f. AV. xvi.4.2, amr'ta stands for 
amr'td u. But even a final -am suffers elision and crasis before u, 
and is misunderstood by the pada (as nominative): thus (as Roth 
suggests) in x.119.13, grho ydmi dramkrto dev'ehhyo havyavd'- 
hanah; 'And home (grhdm u) I go, satisfied, and take along for 
the gods something from the sacrifice.' In like manner, after the 
resolution of pdrdho (anarvd'nam, masculine) into pdrdham u, 
i.37.1, the particle u with abhiprd gdyata is by no means inappro- 
priate, for it is often used in a summons of that kind. Cf. Kuhn, 
Beitrage, iv.197,208. 

Accusative Singular Masculine. 

The regular ending is -m. Next to the N.s.m. it is the most 
frequent form in the R.V., and occurs 6861 times, in connection 
with 1357 stems. For example : imam occurs 130 times; indram 
335; tarn, 509; yam, 259; yajndm, 183 ; vd'jam, 123; sdmam 
212; adhvardm is found 48 times; dcvam, 48; enam, 68 
gdrbham, 56; tydm and tidm, 46; devdm, 65; mitrdm, 53 
vrtrdm, 82 ; samudrdm, 52; sutdm, 62; sil'riam and sil'ryam, 90 
stdmam, 17; hdvam (sometimes neuter), 88. 

In some instances the gender is doubtful : thus dbhogdyam may 
be neuter. Nouns elsewhere neuter are sometimes used as mascu- 
lines, and have been included here: thus drtham (etdm), x.18.4 ; 
51.4,6; cakrdm (dpum), iv.1.3; rdtnam (brhdntam), vi.19.10. 
Asmd'kam may be set down as G.p. of the personal pronoun, or as 
A. s.m. of the possessive. Yipvam occurs 139 times; but I have 
not separated the masculines from the neuters, and have enumer- 
ated them among the latter. The total 6861 would require a slight 
correction accordingly. Vd'ram, vi.13.4 (p. vd dram), has been 
counted as one word. 

The only probable irregularity in the ending of the A. s.m. is in 
x.40.8, yuvdrh vidhdntam vidhdvdm urushyathah, 'Ye protect 
the widower (root 2 vidh) and the widow ;' but vidh has no active 
forms, and this difficulty led Roth to the conjecture that vidh- 
dvdm stood for vidhdvam with metrical lengthening, and he 
translates the two words by viduum cultorem. Herein he is 
followed by Grassmann, who adds that " the metrical lengthening 
frequently occurs where the written text has am" (Wb. 1281 ; 
but Ueb., cultorem etviduamf). In\.50.Q,jdndnisnot singular (Roth, 
L.u. G.d. W.p.82) ; bhuranydntam is to be made the direct object 
of pdpyasi, and not joined to the stock-phrase jdndn dnu. In viii. 
33.8, ddnd' mrgd nd vdrandh purutrd ■ cardtham dadhe, 'To the 
feast, as a wild animal, hither and thither he takes his course,' 
ddnd' mrgd is for ddndm mrgd, unless we assume that ddnd (m.) 
is used once as a neuter plural. 



A.s.m.] Noun-Inflection in the Veda. 331 

Elision and crasis : dpurh nd cakrdrh rdthieva, ' like the swift 
wheel of a chariot (rdthiam), as it were (iva) ;' p. rdthyd-iva, 
iv.1.3. Both texts have -am iva in iv.18.5, where the metre 
demands eva. Compare the usage of the Atharvan (Whitney, 
Pratijakhya, ii.56) j -am iva is contracted to -eva 25 times out 
of 40. 

Nominative and Accusative Singular Neuter. 

The regular ending is -m, without distinction of form for the 
two cases. Nominatives and accusatives s.n. (taken together) 
occur 4275 times, in connection with 949 stems. For example 
antdriksham occurs 45 times ; amr'tam, 31 ; iddm. 165 ; rtdm, 70 
kdm, 48 ; kshatrdm, 31 ; ghrtdm, 47 ; cakrdm, 35 ; citrdm, 34 
drdvinam, 41 ; ndktam, 32 ; paddm, 46 ; pavUram, 37 ; pratham- 
dm, 38 ; priydm, 44 ; bhadrdm, 39 ; rdtnam, 44 ; vipvam (m. and 
n.), 139; sakhydm and sakhidm, 43; satydm, 40; suvi'riam, 57; 
havydm, 33. 

The gender is not always to he determined : as ajakdvdm, 
odandm, kshirapdkdm, nakhdm, muhdrtdm, rdmdm, Id' yam, 
sometimes 2 vdram, vd'ram, pi'pdlam, pd'lam. In v. 87.9, hdvam 
is used as n., although generally m. A few of the occurrences of 
vipvam are m. Ga rdtham, i.70 7; 72.6, has heen counted as one 
word. Some of the words included above are accusatives s.n. in 
form and adverbs in function : as kdm (?), nisvardm, pratardm, 
vitardm, pZ'bham, sapitvdm, sdmantam, sardtham. 

In viii.29.8, prd pravdseva vasatah, we have an example of 
crasis after elision of final m (so Roth). The pada reads pravdsd'- 
-iva. So too we may resolve sdmaneva in iv.58.8 (Nir. vii.17) 
and vi.75.4 (Nir. ix.40) ; cf. x.168.2: and I would even add 
sdmandvagdtya {-am ava-), vi.75.5. Cf. NA.p.n. and Ath. Pr. 
ii.56. Similarly, x.103.11, asmd'kam indrah is to be pronounced 
asmd'k'endrah. 

Instrumental Singular Masculine and Neuter. 

The general case-ending is -d; but the Ls.m.n. of a- stems ends 
regularly in -ena. If this is historically developed out of -a-in-d, 
we might expect to find a good many instances in which the final 
is not shortened ; and in fact, out of a total of 1071 instrumentals 
s.m. and n., there are 85 in -end. Moreover, the still more organic 
form -a-d or -d ought to have left traces. These, again, number 
114. The stem vir'ia exhibits three forms ; arranged in historical 
order, they would stand as follows: *vtria-d (cf. mahitvdd), virid, 
vtriend, viriena. 

1. Of the regular ending there are 872 examples; 374 are m. 
(from 150 stems) and 344 are n. (from 157 stems), besides which 
tena occurs 57, and yena 97 times. In two cases the final vowel 
is nasalized in the samhitd at the endvof a pdda, before an initial 
vowel in the next: i.33.4, ghanenan ekas, and i. 110.5, tejanenan 
ekam. Examples are: mdrena, 34; vdjrena, 33; yajnena,\3; 
sti'riena, 37; ghrtena, 31. 



332 



0. R. Lanman, 



[a-stems. 



2. The following is, I believe, a complete enumeration of all the 
forms with long final (85). 

ix.TO.2. iibhe dyd'vd M'viend vi facrafhe 

ix.80.1. br'haspate ravathend vi didyute 

ix.84.5. viprah kavih kd'viend smrcandh 

x.48.4. purUhinarii sd'yakend hiranyayam 



i.32.5. 


kulicend vivrknd 


vi.32.2. 


su'riend kavind'm 


i.33.13. 


vrshahhend pwro 'bhet 


vii.18.17. 


petuend jaghdna 


i 116.24. 


acivend ndva dyu'n 


viii.85.2. 


vithurend cid astrd 


i.117.21. 


bakwrend dhdmantd 


ix.ee.i'?. 


kd'viend kavih son 


U64.30d 


martiend sayonih 


x.42.10. 


vrjdnend jayema 


1.164.386. 


martiend sayonih 


x.56.3. 


vd'jinend suveni'h 


iii.31.12. 


skdmbhanend jdnitri 


x.68.8. 


viravend vikr'tya 


iii.32.2. 


md'rutend ganena 


x.88.4. 


d'jiend vrndnd'h 


iv.18.5. 


viriend nirshtam 


x.111.2. 


tavishend rdvena 


iv.34.8. 


daiviend savitrd' 


x 139.4. 


tad rtend vi Ayan 


iv.39.3. 


varunend sajoshdh 


x.180.1. 


ddkshinend vds&ni 


AV.iii.12 


.8. amr'tend satnandhi 


ix.10.8. 


martiend sayonih 


v.1.5 


kd'viend krnomi 


ix.10.16. 


martiend sayonih 



RT. vii.55.7, AT. iv.5.1. tend sahasyend vaydm 
AV. vii.60.1, aishydmi Vhadrend saha 

The substantive forms in -end are so few, that m. and n. have 
been given above together. In four instances nd is the eighth 
syllable of a jagati-pdda; in 26, the eighth of a trishtubh; in 
three, the sixth of an anushtubh. In all cases the pada reads na. 
See Rik Prat, viii.21, and cf! Ath. Prat, iii.16. 

Tend, y'end, svend appear invariably in the pada as tena, yena, 
svena. This long final occurs without exception in the second 
place of the verses concerned, of which 13 are anushtubh, 8 jagatt, 
and 7 trishtubh, and is in every case followed by a single consonant 
and that by a short vowel. T'end is first word of an anushtubh- 
pdda 6 times: i.49.2: vii.55.7: viii.20.26; 67,6: ix.61. 19. AV. iv.5.1 ; 
y'end,1 times: i.50.6 : vi.16.48: viii.12.2a; 17.10; 19.20; 24.25: 
x.126.2. Y'end stands at the beginning of a jagati-pdda 8 times : 
v.54.15; 87,5: viii.3.9; 3,i0; 12,1; 12,2c; 12,4: ix.108.4; and 
as first word of a trishtubh-pdda 6 times : i.62.2c (=ix.97.39c) ; 
i.72.8 ; 80.2: ii.17.6: iv.51.4 (18 of the 21 occurrences oiy'end are 
given by Muller, Prat, vii.28). Svend occurs as first word of a 
trishtubh-pdda vii.21.6. For other occurrences in the Atharvan, 
see Whitney, PriLticakhya, iii.16 note, and Benfey, Abh. etc. 
xxi.18,32. 

End, end'. In four instances, where the unaccented Ls. of the 
demonstrative stem a is used as a substantive pronoun, the pada 
reads ena: v.2.11, stiarvatir apd end jayema, 'Let us win by it 
(stdmena) the heavenly waters;' x.108.3, d' ca gdchdn mitrdm 
end dadhdma, ' If he come hither, we will make friends with him 
(indrena).' 1 These two cases are precisely analogous to those of 
the nouns above. For ix.96.2 the Prat, viii.19 prescribes the 
lengthening; vidvd'n end sumatim ydti dcha, ' By it (by ascending 
the wagon) the wise one gets into favor.' In i.173.9, the final, as 



I.s.m.n.] Noun- Inflection in the Veda. 333 

syllaba anceps, remains short in the samhitd/ dsdma ydthd 
sushakhd! ya ena, ' In order that we may be good friends with 
him.' 

With the accented form end' the case is different. The pada 
invariably reads end' . 

It occurs as an adverb or crystallized case-form 17 times, with -d' in both texts, 
and that in one instance even where the samhitd requires a short, vi.20.10 (Kuhn, 
Beitr. iii.123). It stands as follows: anushtubh, i.30.3 and ix.61.11, second place; 
v.73.4, sixth place; trishtubh, x.14.2; i.48.3. second place; i.164.17,18,43 : 
ix.97.52: x.82.5; 125.8, eighth place; i.164.5: ix.97.53: x.27.21; 31.8, fifth place; 
finally x.112.6, where the text looks suspicious. 

As an adjective pronoun it occurs 17 times with its substantive in the same 
pada; and four times it is used substantively: vii.7l.4c, d' na end' naasaty&pa 
ydtarn (sc. rathena); vi.44.17. end' manddno jahi cara cdtr&n (sc. pd'tarena, v.16); 
x.l44.5c,rf, end' vayo vi tdri d'yur jivase, end' jdgdra bandhiiM. (sc. somena). Of 
these 21 occurrences, (a) 12 are such as show the syllable nd' in the second place 
of the verse : jagatlnAH.be; trishtubh iii.34.4: vi.4.417: x.14.4; 85.27; and with 
crasis i.105.19 andii.9.6; anushtubh ii.6.2: v.19.3; 53.12: vii.16.1: x.144.5* (b) 
A long is perhaps favored in the fourth place of a jagati, trishtubh, and anushtubh 
respectively iv.36.7 : vii.71.4, and viii.5.39. (c) The remaining six occurrences are 
in the fifth place: of a jagati ii.34.]4 and viii.21.6 before namasd; of a 
trishtubh i.11 1.1 and vii.93.7, before namasd; viii.85.8, before havishd; x.23.7, 
before sakhid'. 

Aside then from the word end', where analogy led the pada-kdra 
to treat the pronoun in the same way as the crystallized adverb, 
the samhitd shows no forms in -end which cannot be explained on 
the ground of metrical lengthening ; while, on the other hand, 
they may just as well be regarded as relics of the original form of 
the instrumental, whose preservation was favored by their place 
in the verse. 

It is only because there is an d priori probability that the long 
d is organic, that these forms can be taken in evidence at all, as 
explaining the genesis of the ordinary -ena (see Bopp, Vgl. Or. 3 
i.326, § 158 ; and Schleicher, Ootnp.* p.561, § 258). Regarded as 
relics, the uniformity of their preservation in the cadence of the 
verse is remarkable. I have searched diligently every passage 
containing an I.s.m. or n. of an a-stem (1071 in number), and 
have found only one genuine instance where the final is not long 
if required by the cadence. A short is extremely rare, where a 
long is merely favored by the verse: as ix. 109.14, ybna vipvdni, 
better y'end: pankti. So viii.41.10, skarnbh'end. The solitary 
instance referred to is ii. 11.19, vipvd spr'dha d'riena ddsiiin; 
I read vipvd [it] spr'dha d'riena ddsy-dn (?) as i.179.3 : v.55.6. 

If this requirement is satisfied by a double consonant forming 
position, the long final does not appear in the written text. These 
would seelm to be crucial test-cases for this question; for an 
original long ought at least to hold its own in the eighth place of a 
trishtubh. Thus we have iii.31.12, vishkabhndnta skdmbhanend 
jdnitri; but x.111.5, edskdmbha cit Jcdmbhanena skdbhiydn. So 
i.163.3, dsi tritd guhiena vratena; vii.63.4, nUndrh jdndh sU'riena 
prdsUtdh. Again viii.3.9, y'end ydtibhyo bhr'gave dhdn'e hite, but 
y'ena prdskanvam d'vitha. In viii.55.9, kd'd H nil asya d'krtam 
vol. x. 46 



334 C. R. Lanman, [a-stems. 

indrasya asti patinsiam : h'eno nu karii promatena nd pupruve 
jantishah pari vrtrahd' — the -a of prdmatena may perhaps be 
regarded as a syllaba anceps (of. Prat, viii.29). In ii.6.2, ayd! te 
agne vidhema ii'rjo napdd, dpvamishte : end' sflkteua sujdta — the 
trochaic movement would require sdkt'end; but the true metrical 
character of the verse may have been lost sight of. In x. 148.2, 
pronounce ddasir vipah sd'riena sahydh. 

As uniformly and invariably as a long vowel appears when 
required in the cadence, just so arbitrary does it seem to be at the 
beginning of the pdda. I can discover no principle determining 
the quantity of the final of yena or tena in the second place. In 
general yena and tena are of course much more frequent than 
y'end and tend. The third syllable is generally short also when 
yena is used ; so that that seems to make no difference. And 
either form, -na or -nd, seems to occur without distinction as 
regards place in the stanza, at the head of pdda «, b, c, or d. 

Schleicher's I.s.m. with inserted y has no. existence. His only 
example is svdpnayd; but this is a stereotyped adverbial case, of 
feminine form. See d-stems, I.s.f. 

3. The existence of instrumental forms like yajnd' (=. yajnena) 
is unquestionable. I have noted 114 instances. Some will find 
more, and others less, according to their exegesis of certain 



The I.s. in -d is found with m. stems as follows (3V cases) : 
i.168.1, yajnd'-yajnd, 'with every sacrifice ;' and vi.48.1, parallel 
with gird'-gird. Krdndl (cf. a-stems, N.s.m.), ' with longing, 
eagerly:' i.58.3; 139.1 :' v.7.8; 10.2: ix.86.19 (SV. i.559, prdnd'); 
102.1 (SV. i. 51 0, prdnd') : x.61.1. In all seven places the pada 
has krdnd'. See B.'R. v. 1339. Tud', I.s. of the stem tud (in 
compounds only) : tud'-dta, 'holpen by thee;' ii. 11.16, tud'-Htd id 
indra vd'jam agman, 'by thee, O Indra, holpen, came to riches;' 
so vi.16.27 ; tvdti, ' enjoying furtherance (iiti) by thee (£««'),' thrice 
with crasis ; tud'datta, ii.33.2, tud'dattebhi rudra pdmtamebhih 
patdrh himd apiya bheshajebhih, ' with balms most wholesome 
given by thee, O Rudra, may I attain unto a hundred winters :' 
so viii.81.18; tvd'ddta, 'given by thee,' i. 10.7 : iii.40.6 : v. 7.10 ; 
39.1; tvd'hata, 'slain by thee,' vii.32.7 ; tvd'ishita, 'impelled 
by thee,' viii.66.10, with crasis. 

In i.36.16, ghaneva vishvag vi jahi drdvnas, Gr. proposes 
ghanena; but Agni does not slay ' with a club.' The ghand is 
Indra's weapon (i.33.4). I therefore prefer not to change the 
text, but to follow the pada, ghand 'iva, 'As with a mace (as 
though thou didst use a mace), all apart, smite the evil-doers.' In 
i.63.5, ghaneva vajriii chnathihi arntirdn, p. ghand'iva, the 
collocation of words is such that the sense is nearly equivalent to 
tvdrh vdjrena pnathihi, so that iva is by no means superfluous ; 
' as with a mace, do thou with the thunderbolt destroy the foes.' 
And in fact Say. says : vajrivann indra ghaneva ghanena 
kathinena parvateneva vajrena pnathihi. In ix.97.16, the exceed- 
ingly figurative language ghaneva vishvag duritd'ni vighndn, p. 



Ls.m.n.] Noun- Inflection in the Veda. 335 

ghand'-iva, addressed to Soma, makes the iva all the more 
necessary; 'as with a club, all asunder, the dangers driving away 
(flow thou).' Say. takes it as instrumental. Finally in i.8.3, 
where Grassmann ( Ueb. ii.504), reading ghandm, gets the same 
sense as Say. (patrupraharandya atyantam drdham), I consider a 
change again unnecessary (Gr., Wb. 421, ghand's), and interpret 
the ghand! here as I. denoting accompaniment ; ' O Indra, through 
thy favor may we receive (have to use) thy thunderbolt along with 
thy mace (and so) conquer in the fight our foes.' Ddnd', ' with a 
gift,' v.52.14 (Say. havirddnena), 15 ; 87.2. viii.20.14; it may 
also be referred to the stem ddmdn (see man-stems, Ls'.m.) ; for 
ddnd', viii.33.8, see a-stems, A.s.m. In x.25.4, krdtum nah soma 
jivdse dhdrdya camasd'n iva, the pada has camasd'n-iva, misled, 
as it would seem, by the correct avatd'n-iva of pdda 
b. The tertium comparationis is indeed doubtful, but camasd'n 
is probably a nasalized Ls.m. ' Our spirit, O Soma, keep 
thou alive, as by the beaker (thine is kept),' with a slight 
play on krdtum. Grassmann ( Wb.) has assumed for the Vedic 
language a stem ghrnd' upon the basis of four passages: 
iv.43.6, ghrnd' vdyo arushd'sah pari gman; v.73.5, pari vdm 
arushd' vdyo ghrnd' varanta dtdpah; i.52.6, pdrim ghrnd' carati, 
titvishe pdvah. In the first two he takes ghrnd' as A.p.f., contrary 
to the pada, and, as it seems to me, to the sense ; in the third, as 
N.s.f. It will be observed that in each case the verb 
is accompanied by pari and is naturally construed with an 
instrumental. This is surely less forced in the fourth passage, 
i.141.4, d'd id ydvishtho abhavad ghrnd' p&cih, than a N.s.f. in 
apposition to ydvishtho. If they ave all instrumental, it is 
unnecessary to set up a feminine stem, and these four along with 
ix. 107.20 can be referred to the stem ghrnd (as is done by B.R.). 
Cf. <A-stems, I.s.m. and suhdvdn TS. iii.3.il 5 . 

The neuters of this form are more frequent (77). In viii.25.18, 
the third pdda, ubhe dl paprau rodasi mahitvd', cannot be 
catalectic, as the metrical structure of the other stanzas shows. 
The requirements of the metre are satisfied with the least violence 
to the text, if we pronounce mahitvdd (as also perhaps x.96.11). 
And this exemplifies the organic form which all the instrumental 
cases under this head presuppose, and from which they, with the 
following, arise by contraction of -a-d to -d. iv.33.10, ye (rbhdvas) 
hdri medhdyd ukthd' mddanta indrdya cakruh suyujd ye dpvd, 
I translate, ' Te who so wisely, in his word rejoicing, for Indra 
made the pair of bays, the docile;' ukthd' mddantas means 'glad 
of the promise' made them — ' of the word ' fetched them by Agni 
from the gods (i.161.2), namely promotion to divine honors. The 
I. is the ; regular construction with mad in this sense. Medhdyd 
is used onto moivov; they made it skillfully ; but they were also 
shrewd fellows to take up with the proposition (i.161.2); for, as 
the context shows, they were glad to do even more than was asked 
of them : Say. , ukthaih stutibhir madanto harshayantah. Kavitvd' 
x.124.7; kavitvand' viii.40.3 ; taranitvd' i. 110.6 ; mahitvd' i.52.13 ; 



336 C. R. Lanman, [a-stems. 

67.9; 68.2; 91.2; 109.6; 164.25: ii.15.6 : iii.1.4 ; 54.15: iv.16.5; 
42.3: v.2.9; 58.2: vi.29.5 ; 67.3, 10; 68.4: vii.13.2; 20.4; 23.3; 
58.1; 61.4; 97.8; 100.3: x.54.1 ; 55.5 (devdsya papya kd'viam 
mahitvd' dyd' [p. mahitvd' adyd~\ : Roth, mahitvdm adyd, with 
elision and crasis ; but this is suspicious at the end of a pdda, and I 
take it as an I. of accompaniment : ' Behold [herein] the god's 
wisdom coupled with might: to-day etc.'); 56.7; 75.7; 88.9; 
89.1 ; 96.11 ; 121.3, 4. In all these 33 cases (except i.67.9), it is at 
the end of a trishtubh-pdda, which the resolution of the final 
contract-vowel would convert into a jagati. Mahitvand' i.85.7 : 
iv.53.5 : v.81.3 : viii.24.13, all at the end of & jagati-pdda; i.86.9 : 
vi.16.20: viii.57.2: ix.100.9, at the end of an anushtubh-pdda; 
rdthid (cakrena), 'with wagon-wheel,' i.53.9; vtrid i.80.15; 
vrshatvd' i.54.2; vrshatvand' viii.15.2; saJchid' x.23.7a; md'kir 
na end' sakhid' vi yatishus tdva ca indra vimaddsya cdrsheh, 
' May no one cut us off from that friendship, made (of old) 
between thee, O Indra, and thy singer, Vimada.' ii.32.2, md' no 
vi yauh sakhid 1 viddhi tdsya nah, ' Cut us not off from thy 
friendship ; be mindful of it, we pray.' B.R. interpret sakhid! as 
Ls.n., because tdsya is singular; so also Say., Hdsya'' tat sakhitvam 
. . . . ' viddhi ' jdnihi. But this is not conclusive ; the tdsya may 
refer to the whole request md' .... sakhid'. viii.75.1, md' no vi 
yaushtam sakhid'. iv. 16.20, brdhma akarma .... nil' cid ydthd 
nah sakhid' viyoshat, ' So we have made supplication, to the end 
that he may not cut us off now from his friendship.' In x.10.1, 6 
ait sdkhdyaih sakhid' vavrtydm, sakhid' is to be taken as I.s.n. ; 
sakhyd'ni does not occur in the AV., nor sakhyd! as plural. It is 
superfluous ' to turn a friend to friendship ;' and ' guilty friendship ' 
(Gr.) is rather more than the word means. Better 'I would fain 
bring him hither by my overtures of love,' or adverbially, 
' vertraulich.'' The Sama variant, i.340, d! tvd sdkhdyah sakhid' 
vavrtyuh, admits either construction, iv.34.1, ratnadheyd {imam 
yajndm ratnadheyopa ydta) is not A.p.n. in apposition to 
yajndrn, but rather an I.s.n. ; because it refers not to what 
mortals give the gods, but to what the gods give to mortals (cf. 
vii.53.3), and corresponds in usage to ratna-dhd! (i. 1.1, and always, 
of the gods) and rdtnam dhd (iv.15.3 : i.47.1, and often) : ' Unto 
this sacrifice come ye with bestowal of blessing.' v.35.5, sarva- 
rathd! ni ydhi, ' Drive over the foe with all thy host of chariots.' 
x.160.1, sarvarathd' vi hdri ihd munca, 'With all thy chariots, 
here thy bays unharness.' x.141.4, indravdyd' br'haspdtirh 
suhdvehd havdmahe; suhdvd, ' with goodly invocation ;' so A. 
Weber. The interposition of br'haspdtim makes it less natural to 
call it A.d.m. In AV. iii.20.6 (where ed. has misprint, susdvd for 
suhdvd), the interpretation as dual is favored by the parallelism. 
And although suhdvd M is an established idiom, it is better to 
take it as dual in RV. vii.44.2; 82.4 ; 93.1. 

From the Atharvan I have the following forms : mahitvd' iv.2.2 
(Rik x.121.3), 5 (Rik4), 4: v.11.3: vii.79.1 ; 80.1: viii.3.24 (Rik 
v.2.9); 9.2: ix.10.3 (Riki.164.25), 9 (Rik x.55.5) : xii.3.5 : xiv.2.32 : 



I.s.m.n.] Noun- Inflection in the Veda. 337 

xix.49.1 ; in all 13 cases at the end of the pdda. Suhdvd vii.47.1, 
kuhU'th devi'm .... suhdvd johammi; and xi.1.26, r'shin .... 
suhdvd johavimi : here a dual is impossible. In vii.48.1, rdhd'm 
ahdm suhdvd sushtutf huve, the I. is natural, as parallel to 
sushtuti'. But Rik ii.32.4 has the variant suhdvdm, and TS. 
iii.3.11 5 , suhdvdn (cf. camasd'n above). 

There remain finally certain adverbs, which are crystallized case- 
forms: cf. Ktlhner, Grieoh. Gram. 2 i.728: not, etc. Thus and' 
iv.30.3 : viii.21.13; 47.6 : x.94.3, 4: Latin furnishes the Ace. of the 
same stem, enim. Sand iii.54.9: v.75.2: Abl. with adv. accent, 
sand't i.55.2, etc. Of this the stem is used as a real adj. ii.29.3, etc. 
Uccd' 124.10; 28.7; 33.7; 116.22; 123.2: ii.2.10; 30.5; 40.4: ix.61.10: 
x.106.5; 107.2; 183.2: — that is, 12 times; also AV. xiii.2.36. 
It may be regarded as I.s.n. from a stem uced, on account of 
uccais, v.32.6. So nicd' ii.13.12; 14.4: iv.4.4 ; 38.5: vi.8.5 : 
x.34.9; 152.4, on account of nicd't i. 1 16.22, and ntcau AV. ii.3.3: 
iii.19.3 : ix.2.1, 15: v.11.6. It is however more probable that 
nicd' and uccd' are instrumentals of mac and tidac; but, since to 
the apprehension of the speaker the stems were vocalic, uccd and 
nicd, they became the points of departure for the new formations 
uccais, nicd't, and nicais. The same method of explanation would 
require for pa ^cd' (ii.27.11, and 7 times) a stem pas-dnc or pas-dc, 
from which I.s.n. *pas-dcd, paped'. By the side of this, as if it 
were from a stem paped, arose the form paped't (ii.41.11, and 16 
times). 

Dative Singular Masculine and Neuter. 

The D.s. of the a-stems ends in -dya. There is nothing in the 
Veda, to my knowledge, that casts any light on the genesis of this 
difficult form. It occurs 1502 times; 1107 forms are masculine 
(from 310 stems), and 395 are neuter (from 144 stems). In the 
samhitd two of these forms are nasalized: tadvapd'yan eshd, 
ii.14.2, and savd'yan evd', i.113.1. Cf. Rik Prat, xiv.20. 

Among the forms most frequently occurring are the following : 
indrdya, 188; devd'ya,26; mdddya,1§; mdrtidya, 25 ; mitrd'ya, 
23 ; ydjamdndya, 20 ; vdrundya, 23 ; su"rydya, 11 ;—tdnaydya, 
18; tokd'ya, 21 ; sakhyd'ya and sakhid'ya, 29; suvitd'ya, 34. 

Ablative Singular Masculine and Neuter. 

The Ab.s. of the a-stems ends in -dt. There are 389 such 
forms; 183 are m. (from 116 stems) and 206 are n. (from 98 
stems). The latter include some adverbs. In only 389 instances, 
therefore, is the Ab.s. of substantives distinguished formally from 
the G-.s. (for didydt or vldydt, see w-stems, Ab.s.f.). According 
to theory, the organic form is -a-at (e. g., dpva-at). Justi (JTand- 
buch, p. 359, §13) gives 14 such Zend forms (as daSvdatca); but 
Geldner (Metrik des jilngeren Avesta, § 30) says that' the Ab. 
ending -dt counts regularly for one syllable. For the Veda the 
existence of forms in -a'at is extremely doubtful (cf. Kuhn, 
Beitrage, iv.181) : they have been proposed for i.30.21 : viii.5.31 : 



388 0. R. Lanman, [a-stems. 

x.22.6, pardkdat; x.158.1, antdrikshaat; viii.11.7, sadhdsthaat: — 
but in every case the pdda is one of seven syllables with 
catalectic close, and it is far from certain that the fuller cadence 
is necessary. The like is true of caritrdt, viii.48.5. 

Examples of ablatives in -dt are : indrdt, 8 ; updsthdt, 9 ; 
samudrd't, 15 ; — antdrikshdt, 12 ; ddrd't, 19 ; papcd't (adv.), 26. 

Genitive Singular Masculine and Neuter. 

The G.s. of a-stems ends in -asya. There are 3340 such forms ; 
1890 are m. (from 503 stems), and 695 are n. (from 175 stems). 
Moreover asya occurs 402 times ; tdsya, 83 ; ydsya, 203 ; and 
vipvasya, 67. In x.60.2, bhajerathasya has been counted as two 
words (rdthasya). In one instance the final vowel is nasalized at 
the end of a pdda: viii.89.5, rtdsyan ikam, cf. Prat, ii.31. 

Vocalization of the y of the ending is a thing of at best 
doubtful existence. In x.23.6, vidmd' hi asya bhdjanam indsya, 
and Val. 11.4, saptd svdsdrah sddana rtdsya, a catalectic 
jagati-pdda with the penultimate long by position, is certainly 
better than the cadence with resolution, indsia, rtdsia. In 
i.61.13, turdsla kdrmdni ndvya ukthaih (so Gr.), the metre 
is quite out of order ; if we read turdsya .... ndeiya (cf. 
i. 105.15), the caesura is wrong. In x.H4.105(Gr. wrongly rdthasia), 
the sixth syllable is syncopated. The vocalization is possible in the 
case of asia smdLasid: vA5.9,yddasia wrviyd' dirghaydthe; x.61. 
24c, saranitir asia silnur d$vo; xA42.5,prdtiasia prenayo dadrpre 
(better, perhaps, asya prdyinayo) ; in i.23.24, if an extra syllable is 
needed at all, it should be devaas, not asia (see a-stems, V.p.m.) ; 
i.131.6, d' me asia vedhdso ndviyaso; and ix.98.8, asid vo hi 
dvasd. The final a suffers protraction (purely metrical, Benfey, 
SV. lx.) as follows: i. 162. 19, ekas tvdshtur dpvasid vipastd' (cf. 
VS. xxv.42; TS. iv.6.9 3 :— TPr. iii.8,'p. 91); vii.79.4, yd'm 
tvd jajntir vrshabhdsyd rdvena; AV. i.12.2, yd dgrabhit pdrva 
asya grdbhitd; see Ath. Pr. iii.16. 

Examples of this case are: asya, 119; indrasya, 123; devdsya, 
60 ; yajndsya, 55 ; sutdsya, 53 ; s&'riasya and sU'ryasya, 93 ; 
sdmasya, 88 ; — amr'tasya, 35 ; asya, 55 ; rtdsya, 187 ; bhtivanasya, 
39. 

Locative Singular Masculine and Neuter. 

The case-suffix is -i, as with consonant stems. This, united 
with the thematic a, forms -e. There are 2491 such locatives in 
-e; 1235 are m. (from 373 stems), and 1256 are n. (from 309 stems). 
In some instances, where the ending appears in the samhitd 
(Prat. ii. 1 0) as -a, before vowels, the padakdra mistakes them for 
nominatives, and writes them with visarga : thus in iv.17.14, a" 
krshnd tra juhurdnd jigharti tvacd budhn'e rdjaso asyd ydnau, p. 
krshndh; we must interpret, '(Misleading) craftily he darts it 
(cakrdm sil'ryasya) to the black abyss of night, to the depth 
of this sky,' i. e., to the place where the sun remains from its 
setting until its rising. The reading krshne tvacds budhn'e is 



L.s.m.n.] Noun- Inflection in the Veda. 339 

confirmed by the following gloss, dsiknidm ydjamdno nd hotd 
(Say., pHrvarcd saha drshtdntatvena sambadhyate). In vii.76.3, 
jdrd ivdcdranty tisho, p. jdrdh-iva dcdranti, we read jdre-iva: 
' Ushas, as if hastening to her lover.' In i. 112.17, agnir nd'didec 
citd iddhd djmann d', p. citdh, we interpret, ' Was glorious, as 
the fire kindled on the altar (cite iddhd), upon its path.' In 
v.52.10, d'pathayo .... yajndm vishtdrd ohate (3 pi.), the pada- 
text reads vishtdrdh; I understand it thus : Let their customs 
carry them where they may, yet when I sacrifice ' they wait ' 
quietly ' on the straw (vishtdrd) ' for it. Finally, it seems 
necessary to read ibhe suvrate, ix.57.3 (where the text has tbho 
rd'jeva suvratdh) ; 'As a king amid his trusty following.' 

Examples of this case are: adhvare, 68; abM'ke, 22; indre, 33 
updsthe, 49 ; grhe, 23 ; jdne, 26 ; ddme, 40 ; mdde, 48 ; yajne, 28 
sitte, 53; — dgre, 43; antdrikshe, 23; durone, 31; pad'e, 33 
mddhye, 29 ; vlddthe, 49 ; sddane, 35. 

Vocative Singular Masculine and Neuter. 

This case is like the stem in form. The accent, if there be any, 
is always on the first syllable. The vocatives s.m. number 2498 
(from 259 stems). 

There is hardly a single indubitable example of a vocative s. 
neuter in the Rig-Veda. In i.44.5, stavishyd'mi tud'm ahdm 
vipvasydmrta bhojana, 'Immortal delight of every one,' there 
may be two, amrta and bhojana; but it is possible to take bhojana 
as V.s.m. of an active verbal adjective (see Gr., Ueb.) ; or (as B.R. 
suggest, v.388) we may write amrtabhojana, and take it as V.s.m. 
of the stem amrtabhojana (see Lindner, Altindische JSTominalbil- 
dung, p. 42). In v.46.25, pdrdhah prd yanta md'rutotd vishno, 
the sense seems to require us to take md'ruta as vocative with 
$drdhah. For the accent, cf. X.86.13& The 'A V. has several voca- 
tives s.n. They are antariksha vi. 1 30.4 ; traihahuda and devdnjana 
xix.44.6; talpa xii.2.49; and visha iv.6.3, where the MSS. have 
vishah. 

In vii.25.4, the samhitd shows a nasalized vowel before Skas: 
vipved dhdni tavishiva ugran. Cf. Prat, xiv.20. So in viii.15.3 
11, purushtutan eko. 

Protraction occurs in some instances : i.61.1 6, evd' te hdriyojand 
suvrkti; viii.4.1, simdpuri),' nr'sMto asi d'nave (cf. Pr&t. vii.17); 
viii.45.22, abhi tvd vrshabhd sute, and 38, evd're vrshabhd sut'e. 
The pada-text has hdriyojana, sima, and vrshabha. B.R take 
simd as adv., ' allenthalben.' The stem is sinid, and for adverbial 
retraction of the accent I can give no example. The pada-texts 
ofRV. and SV. (i.279, ii.581), and the accent make against the 
view of B.R. Grassmann, in his note to i.6.3— p'epo maryd 
apepdse, p. marydh— takes maryd as V.s.m. The sense requires 
it, and the above examples seem to justify it. Kuhn (Pdli-gram. 
p. 71) refers to this lengthening. I think it is purelv metrical. 
If the requirements of metre are satisfied by a double "consonant, 
the -a remains short: thus, rnahina prdyishthdh vi.26.86; so 
vii.24.15. ' ' ' 



340 C. R. Lanman, [a-stems. 

In the phrase vd'yav indrapca, i.2.5, 6, indras is taken by 
Grassmann as a V. ; similarly i.18.5; 135.4: iii.25.4 : iv.47.2, 3: 
vi.69.8 : vii.97.10; 104.25: ix.95.5; the peculiarity is, however, 
doubtless a syntactical rather than a formal one, and the case a 
nominative. 

The stem indra is one of the few, complete sets of whose forms occur in the s. 
It occurs as follows : N. 523; A. 335 ; 1.34; D. 188; Ab. 8 ; G. 123; L. 33; 
V. lOOt ; N.A.V.d. 11 ; in all, 2262 times. 

Examples of vocatives are: amrta, 12 ; ugra, 23 ; deva,\Z2; 
puruhilta, 49 ; pavamdna, 63 ; yavishtha, 29 ; mitra, 35 ; varuna, 
45 ; vrshabha, 27 ; ptira, 94 ; s6ma and soma, 240. 

Nominative Accusative and Vocative Dual Masculine. 

The Vedic ending of the N.A.V.d.m. of a-stems is d. In the 
Rik the ending au is exceptional (and by no means organic ; cf. 
Benfey, Gram. p. 303). The comparative statistics will be given 
elsewhere. As regards the circumstances of occurrence of these 
endings a few interesting observations may be made. 

A. a occurs in the Rik 1129 times. It is the ending used 

1. at the end of a pdda/ 

2. before consonants; 

3. before an initial vowel with which it is fused ; 

4. It is never used before vowels with hiatus. 

B. au occurs 171 times. It is the ending used chiefly before 
vowels, where it forms a separate syllable dv, without hiatus. 

The determinant of the form is therefore to a certain extent the 
metre. The regular form is d, and it is used before consonants, 
and before an initial vowel where the words could not be pressed 
into the metrically limited verse save by its fusion with that 
vowel ; otherwise, hiatus not being tolerated (exceptions below), 
dv was used before such an initial. These differences are 
illustrated by i. 184.1, td' vdm adyd td'v apardm huvema, and 
i.14.3, mitrd' gnim pushdnam bhdgam; so too most strikingly by 
i.93, passim: e. g. dgnishomdv imam, but dgnishomd yd. 

Of the 1129 d-forms: 

1. 230 are at the end of a pdda : 62 at the end of pdda b, and 45 
at the end of d or of the verse ; 81 at the end of a, and 42 at the 
end of c. In 26 of the 81 cases, pdda b begins with a vowel, and 
then the two are always written as fused in the samhitd; but 
(unlike the epic cloka) they never coalesce in reality : thus, i.28.2 
a,b. vii.66.17 a,b. So in 8 of the 42 cases, as i. 11 6.10 c, d, dasrd 
d't: samhitd-text, dasrd' d; 118.3c, a"; vi.62.2c, d: See Bohtlingk, 
Chrestomathie* , p. 342. There are only two instances, on the 
other hand, of au in the pause b, d: ii.30.6 b, codau, and v.47.3 d, 
dntau. 

2. 799 stand before consonants (in the interior of a pdda). 137 
are before v, and 141 before other labials ; 126 before sibilants, and 
395 before other consonants: i. e., 278 labials against 521 conso- 
nants of the other classes. The frequency of v is significant as 



N.A.V.d.m.] Noun- Inflection in the Veda. 34L 

regards the development of au from d. Cf. to! vdm and td'v 
aparam; and see below. 

The ending au occurs in the Rik exceptionally before consonants 48 times : 
i.179.6: ii.27.15; 40.1, 2 quater : iii.33.13; 53.17: iv.41.2: v.34.8 bis; 36.6 ; 
40.7 ; 66.1 ; 68.4: vi.59.5; 60.14; 74.4 bis: vii.50.2; 84.2: viii.33.19 bis; 35.4, 5, 6: 
ix.112.4: x.14.10 ter, 11 bis, 12 bis; 27.20 bis; 85.11, 18,42; 113.7; 111.9 bis; 
136.5; 137.2 bis; 178.2 bis. That is, in only 27 different hymns; before labials 
17 times, sibilants 15; 5 times at the end of pdda a, vi.74.4: x.14.12; 27.20; 
85.11, 18, and once at the end of c, ix.112.4. An examination of these passages 
yields striking results, as proving how the criteria of dry grammatical investigation 
confirm the results of criticism based on other methods. The dual-form in au 
is found either in the last verse of a hymn (i.179.6: iii.33.13: v.36.6: vi.74.4 
bis: viii.33.19 bis: ix.112.4: x.117.9 bis), or the last but one (v.34,.8 bis; 68.4: 
vi.60.14: x.178.2 bis), or some verse near the last, which has been patched on 
(x.27.20) or interpolated (vi.59.5) by later hands; or the whole character of the 
hymn betrays its later origin (as vii.50 : x.85; 117; 136; 137; 178). And it is 
something more than a mere coincidence that 1 5 of all the instances of -au before 
a consonant, about one-third, occur in verses found in Grassmann's Arihang : 
Vebersetzung spdter angefiigter Verse und Lieder. 

3. 93 coalesce with a fpllowing vowel (in the interior of a pdda) : 
53 to e, as v. 76.2 b, apvinopastutehd; viii.34.96 and often with 
iva; 24 to d, as i.161.7 td'krnotana; 182.4 d: v.46.3 a; and 16 to 
o, namely i.34.9; 36.17; 116.10; 117.1; 183.5: ii.39.8 : iv.14.1 ; 
21.9, bhadrd' te hdstd sukrtotd pdni' '; 44.4: vi.48.4 : vii.63.5 ; 
64.2; 71.4: Val. 9.4: x.87.3*; 125.1. ' 

4. The exceptions to A. 4 are as follows : 

a. The cases in which a is followed by u- or u- in the interior of a pdda form 
a class by themselves. In the 16 passages just cited, the concurrent vowels are 
run into one syllable. The Atharvan and later sandhi would write -dv u-: thus 
AV. iv.32.7, ubhd'v updnpu; so x.4.8, td'v ubhd'v arasd'; xi.2.14, samviddnd'v ubhd'v 
ugrau; TA. iii.12.5, kd'v thru'; but once, xx.136.?, rnushM' upd'vadhit, like the 
Rik. If the metre requires them to be kept apart as two syllables, the Rik- 
samhitd writes them with hiatus, and the padapdtha reads always -au u-. (Auf- 
recht", pada-extr&cts to x.83.7, a mere omission ?) The cases number 1 5 and are : 
i.2.9 = SV. ii.199, tuvijdtd' urukshayd; 13.8, sujihvd' upa; 35.6, dvd' updstJid; 
iv.41.10, cakrdnd' utibhis; v.65.3, pu'rvd upa; viii.22.13, id' ■&; 14, td' ushdsi ; 
40.3, td' u; 1 6.4, vdvrdhdnd' upa ; 90.2, varshishthakshatrd urucdkshasd ; x.83.7 
(=AV. iv.32.7, ubhd'v), ubhd' updnpu; 90.11, kd' uru'; 93.6, mitrd'varund 
urushyatdm ; 106.1, ubhd' u; and once in vii.70.4, devd oshadhishu, p. devau. 
(x.90.11 = AV. xix.6.5, kirn uru'; VS. xxxi.10, kirn uru,'; TA. iii.12.5, kd'v uru'.) 
This fact is significant, when taken in connection with what was said above about 
the labials. At the end of the pdda, the two vowels are fused to o in the 
samhitd, but are of course always to be read with hiatus: i.22.2, yd' swathd 
rathi'tamobhd' devd' divispr'cd, read -d u-; i.93.6: v.64.4: viii.9.9; 61 17 • 
x.106.4. 

/3. Of the 1129 forms there remain seven. Six seem to precede a vowel with 
hiatus. As in the case of the Rik <ro-forms before consonants, the exceptions 
here only confirm the rule, v.41.3, d' vdm yaishthd apvind huvddhyai; the hymn 
is not homogeneous, and abounds in false readings, x.132.2, sushumnd' ishitatvatd 
yajdniasi; the metre is in the utmost confusion, alike in no two consecutive verses, 
and the sense is obscure, v.65.6 (last verse), yuvam mitfa)remam janam, perhaps, 
as in v.40.7, tuam mitaro. vii.70.1, d' vicvavdrdfvind gatam nah; perhaps it once 
began ehd vif- or d' vicvavdrdv acvind gatam nah. In x.22.5, perhaps we have to 
read d' agd rjrd' (a)tmand vahadhyai ; atmand is the link between dtmand and 
tmdnd (accent I); but the metre is really hopeless. In vi.63.1, kua tyd' valgu' 
puruhutd' adyd, the hiatus is not the only trouble ; we need a short, puruh&td (see 
below). 

vol. x. 47 



842 0. B. Lanman, [a-stems. 

y. In vi.67.1, finally, dud' jandil asamd bdhubhih svaih, the hiatus is prevented by 
nasalization (but p. j&ndn). So too i.35.6, upasthdti | iM. 

(5. Some apparent cases of hiatus our canon teaches us to avoid, as follows : 
pronounce, i.36.17, agnih pro, dvan mitrota medhydtithim ; ill 0.2, uta arushd'ha 
cakre vibhrtrah (not utd'rushd' aha); vii.50.1, d' maam mitrdvarunehd rakshatam 
(not -nd iha) ; vii.67.5, prdaclm u devdcvind dhiyam me. 

B. 1. In the Rik -au occurs 171 times. In the older portions 
thereof, it is used only before vowels, and is written -dv. This is 
the case in 121 instances, 70 per cent, of the whole number. Its 
occurrence before a consonant may even warrant a suspicion of 
the lateness of the verse concerned (the 48 cases are above, A. 2). 

2. Of the 388 cases in which -au occurs in the Atharvan, in only 
102 does -au stand before a vowel (as -dv) ; i. e., about 26 per 
cent., against 70 in the Rik. 

3. The Atharvan has -au as varia lectio in some passages where 
the Rik has -d: iv.32.7, ubhau = x.83.7, -d'; vii.58.1, dhrtavratau 
= vi.68.10, -a; viii.3.3, ddnshtrau = x.87.3, -d; in xviii.3.12, 
mitrd'vdrund, the Ath. pada has -nau. 

There are 47 forms in -d in passages peculiar to the Atharvan 
(not found in the Rik). They are distributed as follows : at the 
end of a pdda, 7; before consonants, 33 (of which 7 before v) ; 
coalescing, xiv.2.64, to e; iii.4.4 and x.4.16, to o; making hiatus, 
vi.3.3: xix.13.1: xx.129.15, 16. 

The N.A.V.d.m. appears with shortened final as follows : 

1. Cases where the pada reads -d; the samhitd has -a : 

a. For the sake of the metre, in the seventh place of ajagati: i.151.4, prdsd' 
kshitir amra yd' rnahipriyd'; of a trishtiibh: vi.68.5, indrd yd vdrh varuna dd'fati 
tm&n ; vii.61.1, ud vdrh caJeshur varum supratikam ; in the fifth place of a gdyatri : 
i.15.6, mitrdvaruna duldbham (p. -nd); 17.3. indrdvaruna rdya d' (p. -nd); 17.7, 
indrdvaruna vdm ahdm (p. -nd); 17.8, indrdvaruna nu' nu vdm (p. -nd); 17.9, 
indrdvaruna yd'ih huve (p. -nd) ; v.67.1, balitthd' deva nishkrtam (p. devd) ; /3. at the 
end of a gdyatri-pdda (metre indifferent), i.15.6, yuvdrh ddksham. dhrtavrata (p. 
-vratd); v.64.6, yuvdrh no yeshu varuna (p. varund); y. against the metre v.66.6, 
mitra vaydm ca surdyah (p. mitrd); vii.60.12, iydrh deva purohitir yuvdbhydm 
(p. devd) ; 85.4, yd dditya cdvasd vdm ndmasvdn (p. ddityd) ; viii.9.6, ydd vd deva 
bhishajydthah (p. devd). These are all given in the Pr&t. iv.39,40. 

2. Cases where both texts read a: v. 74.4, paura paurd'ya 
jinvathah (p. patira) ; vi.63.10, bharddvdjdya vira nti! gire ddt 

(p. vira); ashtd, x.27.15, seems to be a shortened dual-form (cf. 
i.35.8, ashtaii), like the later dub, ambd, 6vo. Compare also 
indrdvdyu' i.2.4, and mitrdrdjdnd v. 62.3. 

3. A merely graphic peculiarity is the short final a at the end 
of an odd pdda, where the next begins with r-: ii.3.7, vidushtara \ 
rj'd, p. viduh-tard; vi.68.2, tuvipushrna \ rtena, p. -md; viii.66.11, 
's4samskrta'\ rdtipe, p. td; x.66.13, piirdhita | rtdsya, p. -td. 

4. The metre demands a short in vi.63.1, Ma tyd' valgti! 
puruhutd' adyd : see above ; and vi.67.8, tad vdm rnahitvdm 
ghrta-anndv astu. Cf. Kuhn, Beitrdge, iii.120-121. 

i'.112.18, yd'bhir angiro mdnasd niranydthah. For angiro as 
a genuine dual-form there is no support. See Gr., Wb. 14, 
and Kuhn, Beitr. iii.121. Because the Prat, allows -o for -au in 
the Loc, we may not put -o for the proper dual-ending -d. We 



N.A.V.d.m.] Noun-Inflection in the Veda. 343 

might explain it as an instance of crasis, for angird u, 'And with 
what helps, ye two Angiras ;' but there is no support for giving 
the Acyins that name. See therefore as-stems, A.p.m. 

It must be remembered^that the above statistics refer to forms 
from a-stems only. The forms from non-a-stems will be found to 
yield substantially corresponding results. The complete enumera- 
tion of the Rig- Veda forms in au follows : 

Nominatives: ansau, aghniau, ajarau, dntau, drdhau, -itau 2, imau 3, 
indrdvarunau, udumbalau, ubhau 4, ur&nasau, rjrau, rshvau, etwk 2, oshthau, 
karnau, kagaplakau, kulphau, kshdyau, gaurau 3, caturakshau 2, codau, jd'yamdnau, 
jdtau 3, jdnau, jushdnai,, jrayasdnau, tau 13, tigrnd'yudhau, dahsishtau, d'Utau, 
devau 12, -hitau, ndvyau, pddakau, punarmanyau, purohitau, prchdmdnau, 
prdyishthau, bhu'ripapm bhedau, mdhikshatrau, mitrd'vdrunau 4, mithunau 2, 
modamdnau, yau 8, ydtamdnau, yamau, yuktau, rakshamdnau, rohitau, raudrau, 
vdrnau, vd'tau, vdvrdhdnau, caphau, cabdlau, fukrau, cyenau, pvetai, s&camdnau, 
sajoshau, samau, samudrau, sdmanau, sdrameyau, sudhdnau, suvi'rau, sucevau, 
sldnau, -stutau, sthirau, hansom 4, harinau, hdstau 2. 

Vocatives: dgniparjanyau, agnishomau 4, indrdvarunau, indrdvarunau 3, 
indrdsomau, ghrtaannau, dasrau 7, devau 9, dhishniau, ndsatyau 4, bhuranau, 
mitrdvarunau 9, rudrau, viprau, cundstrau, somdrudrau. 

Nominative and Accusative Dual Neuteb. 

This case ends in -e, which appears to be the result of fusion of 
the thematic vowel with the general ending 4. The form occurs 
57 times, in connection with 32 stems. There is no instance of a 
vocative. JVdMa, although neuter, is declined as a masculine, and 
used with a feminine adjective, in the combination ushd' sd-ndktd. 
In ii.39.4, yug'eva ndbhyeva must be resolved as yug'e va ndbhye 
va, although the pada has in both instances -d-iva. 

Enumeration : acakre, apdre 2, ime, ubhe 1 5, kramane, cakre 3, te, tanaye 2, tigme, 
tote 2, dvate, kite, nave, pade, pavdste, pdtalye, priyatame, md'Hccaive, ye, rejamdne, 
vddhatre, viddthe, vishurtlpe 2, pirshe, fdshe, cr'nge 5, satydnrte, samante, 
sdcandnapane, sudine, sumeke 2, tastabhdne. 

Instrumental, Dative, and Ablative Dual. 

The ending is -bhydm, before which the thematic a is lengthened. 
There are only 24 of these forms, all told; 19 are m. (from 14 
stems), and 5 are n. (from 4 stems). The oblique cases of the dual 
being so rarely used, the consciousness of their several meanings 
was not definite and clear ; at any rate, a certain relation was not 
always expressed by the properly corresponding form. Thus 
td'bhydm, ix.66.2, seems to be an I. with the meaning of a L. A 
like uncertainty existed as regards the form of some of the 
nominatives and accusatives p.n. See mant-stema. 

There is no evidence that these cases were ever formally 
separated. The exegesis groups them as follows : 

Instrumental masculine : td'bhydm, vi.57.3 : x.137.7; ddcacdkhdbhydm, x.137.7; 
mitrd'vdrundbhydm, v.51.9; yuktd'bhydm, vi.23.1; cubhrd'bhydm, i.35.3 ; hds- 
tdbhydm, x.137.7; haryatd'bhidm, viii.6.36: — neuter: rksd'mdbhydm, x.85.11; 
114.6; ubhd'bhydm, ix.67.25, is joined with two nouns, one m. and one n. 

Datives masculine : dbhyd'm, ii.40.2; tveshd'bhydm, Val. 9.5 : nd'satydbhydm, 
1.116.1 : v.77.4; nicird'bhydm, i.163.1 ; yd'bhydm, viii.38.10; ndasatydbhidm, i.20.3. 

Ablatives masculine: dnsdbhydm, x.163.2; dbhydm, iv.32.22; kdrndbhydm, 



344 C. R. Lanman, [a-stems. 

x.163.1; td'bhydm, x. 14.11: — neuter: indtasnabhydm, x.163.3; prdpaddbhidm, 
x.163.4. 

I add a f ew i orms from the Atharvan : m. kdrndbhydm, ix.4.17 ; 8.2; ddnshtrd- 
bhydm, x.5.43 ; dakshinasavyd'bhydm, xii.1.28; vrlhiyavd'bhydm, x.6.24; sur- 
ydcandramdsdbhydm, vi.128.3 : xi.3.34 ; n. pdrfvd'bhydm, ii.33.3. 

Genitive and Locative Dual. 

The general ending of this case is -os; but between this and 
the thematic vowel of the a-stems there is an inserted y. There 
are 57 of these forms in -ayos; 27 are G.d.m. (from 12 stems), 
and 27 are L.cLm. (from 15 stems) ; besides, tdyos, n., occurs 
thrice. 

Here, too, there seems to be some confusion of form and 
function. Thus dhvasrdyos, ix.58.3, does the duty of an Ab. Cf. 
the preceding section, and «°-stems, G.L.d.m. 

Roth explains avds, vii.67.4, as equal to ayds, ' of these twain ;' 
so vi.67.11, and x. 132.5 (avdr vd[ni]). 

The Zend usually has the inserted y; as, saptayd, Y. lvii.31 ; 
pddhaydo, Yt. x.23 ; but not always; as, fratirdo, baeshataptirdo, 
Yt. xiii.125. The Veda also shows a fe,w such peculiar forms, in 
which the ending -os is added directly to the stem after this has 
dropped its final a. In i.136.1, 5 : vi.69.8 : vii.103.4, we have enos 
(for enayos) ; x.105.3, yds (for ydyos) ; so in x.96.10, pastios (for 
pastiayos), and ix.l02.2,^>asA«o* (for pdshiayos), unless we assume 
a genuine feminine formation for each of these two passages. In 
iii.55.2, purdnioh sddmanoh, we have an anomaly, the reverse of 
that in i.160 i, sujdnrnani dhishdne. 

Genitives d.m. : ayos, vi.25.6 (pronounce dyos); indrdvarunayos, i. 17.1 ; ifdnayos, 
vii.90.5; ubhayos, i.120.1: vi.25.6; tdyos, i.11.6; 21.1; 22.14; 136.3; 164.20: 
v.86.3 ; devayos, vii.61.1; mitrdyos, vi.51.1 ; mitrd'vdrunayos, x.130.5; ydyos, 
iii.50.2: vi.60.4: vii.65.1: viii.10.3, 4; 40.4: x.22.5; '65.5 bis: Val. 11.2; 
yamdyos, x.117.9; vdrunayos, vi.51.1; vaikarndyos, vii.18.11. 

Locatives d.m.: dnsayos, v.57.6; dfvayos, vi.47.9; ayos (p. wrongly dyos), 
iii.54.2 : x.105.4, 9 ; updkdyos, i.81.4; tdyos, x.114.1 ; tuvijdtdyos, vii.66.1 ; naddyos, 
x.105.4 ; mitrdyos, vii.66.1 ; mushkdyos. x.38.5 ; ydyos, ix.58.4 ; vdrunayos, vii.66.1 ; 
vahislithayos, vi.47.9; vivratayos, x.105.4; sthurdyos, vi.29.2 ; hdstayos, i.24.4; 38.1 ; 
55.8; 81.4; 135.9; 162.9 ; 176.3: vi.31.1 ; 45.8: ix.18.4; 90.1. 

Genitives d.n. : tdyos, iii.55.11, 15 : vii.104.12. 

From the Atharvan: kdrnayos, vi.141.2 : xix.60.1; ddnshtrayos, iv.36.2: 
xvi.7.3, etc. 

Nominative and Vocative Plueal Masculine. 

There are two endings, -dsas and -as. The forms in -dsas (N. 
936 -|-V- 101 = 1037, from 403 stems) are about one-half as 
numerous as those in -ds (~N. 1954 -f V. 226 = 2180, from 808 
stems) in the Rik. The Atharvan has in all only 91 forms in 
-dsas (from 54 stems), against 1545 forms in -ds (from 576 stems). 
Deducting those that occur in Rik-passages, the Atharvan has in 
verses peculiar to itself only 57 forms in -dsas against 1366 in -ds. 
That is, the ratio has changed from 1 : 2 to 1 : 24. The details of 
comparison will be given hereafter. In general, accordingly, the 
ending -dsas is the older, and goes gradually out of use in the Vedic 
period. On the other hand, we may not stretch the interpretation 



N.V.p.m.] Noun- Inflection in the Veda. 345 

of these figures. Both forms stand often side by side, as in iv.25.8, 
and I believe that the choice of the poet between the longer and 
the shorter form was often decided simply by the requirements of 
the metre. This is illustrated by the comparison of some similar 
verses: v.59.6, t'e ajyeshthd' dkanishthdsa udbhidah, but v.60.5, 
ajyeshthd! so dhanishthdsa ete; so ii. 1.16, brhdd vadema viddthe 
suvt'rdh, but ii. 12.15, suvfrdso viddtham a" vadema; and by the 
Atharvan variants of Rik passages: AV. iv.31.1, hdrshamdnd 
hrshitd'so marutvan, RV. x.84.1, hdrshamdndso dhrshitd l 
marutvah; AV. v.2.4, md! tvd dabhan durevdsah kapdkdh, RV. 
x.120.4, md'tvd dabhan ydtudhd'nd durevdh; AV. v.3A,vipve devd' 
abh'i rakshantu mehd, RV. x.128.4, vi$ve devdso ddhi vocatd 
nah; AV. xix.13.11, asmd'n devdso avatd hdveshu, RV. x.103.11, 
asrnd'n u devd avatd hdveshu. The difference between RV. iii.29.9, 
yena devd' so dsahanta ddsyiln, and AV. xi.1.2, yena devd' dsahanta 
ddsyiin, rests perhaps upon a mere haplographia. 

These comparisons suggest a most plausible remedy for some 
defective pddas. I have found no cases of redundant pddas 
needing the substitution of -ds for -dsas. On the other hand, 
pddas deficient by one syllable are not uncommon, and the 
emendation (-dsas for -ds) seems certain in several cases : 
iv.37.4, dyahpiprd[so] vdjinah sunishkd'h; v.41.9, tuje nas tdne 
pdrvatds[ah\ santu; x.94.11, trdild'\so~\ dtrdildso ddrayah; also 
AV. xi.1.2; RV. vii.35.14^= x.53.56, g6jdtd\sa\ utd yeyajhiydsah 
(Gr. suggests -tads, but a is frequent in the third place, as 
in i.89.16: ii.27.3 5: iv.4.12 5); x.78.6, $ig,i%'ld\so\ nd krtddyah 
sumdtdrah (Gr., -aas). In v.41.4, djim nd jagmur dpvd'fvata- 
rnds\ah~\, dpuapvatdmdh is possible. In vii.56.3e, vd'tasvan&sas is 
formally justified by iv.6.10, where the stem tuvishvands occurs; 
the pentasyllabic metre, however, requires the form from the 
vocalic stem, vd'tasvandsas (of. x.46.7 a, c). 

On the contrary, in vii. 9*7.6, tarn [u] fagmd'so arushd'so dpvdh, 
the insertion of u (cf. verse 3) seems to me better than either the 
resolution arushdaso (Kuhn, Beitr. iv.183), or the emendation 
dpvdsah. So in i.163.10 and iii. 8. 9, $rayini$ds is better than the 
quite possible hansd'sas. Roth suggests sidmatashtdras for -tdsas, 
x.15.9. ' ' " 

Grassmann (Wb.) suggested the resolution of d to aa or ad in 
x.77.2, divds putrd'sa, etd nd yetire; 78.2, supdrmdno nd, sdmd 
rtdrn yate, and 78.3, pitfndam nd, pdnsdh surdtd'yah. These 
pddas are peculiar in structure and movement. Their rhythmic 
accent is as follows : ~ -' ~ -' ~, ^ -' ~ -' _ ^. I n each, the syllable 
ds is the first accented one after the caesura, and requires no 
resolution (Grassmann, TJeb. ii.479). The final pdda of ii. 1 ; 2; 
11 ; 13 ; 23 ; 27 ; 33 ; 39; 40; 42; 43 : ix.86, brhdd vadema viddthe 
suvt'rdh, can be read with catalectic cadence, since it is a stereo- 
typed refrain, and need not be made to correspond with the other 
pddas of its verse by the resolution suvt'raah. 

The resolution aa in devaas, i.23.24: vii'.66.2: x.137.5 • 159.4 • 
174.4, and in dd'naas v.27.5, is in each instance at the end of an 



346 0. R. Lanman, [a-stems. 

octosyllabic pdda, and of doubtful necessity ; but in devaas, i. 105.5, 
it occupies the fifth and sixth places. See p. 338, top. 

In L127.7 and vi.67.10, kistd'sas ought to be read as a dactyl. 
Perhaps it is an erroneous way of writing the original kistd's, 
which was pronounced Mstdas. If so, darpatd'sas ix.101.12 is an 
analogous instance. 

Examples of the forms in-«sasare: amr'tdsas, il; ddityd'sas, 
24; jdndsas, 41; devd'sas, 86; yajmydsas, 21; sutd'sas, 29; 
sdmdsas, 41: — of those \n-ds: amrtds, 22 ; ddityd's, 39; jdnds, 
24; d'evds and devds, 311 ; yajniyds, 10 ; sutd's, 27 ; sdmds, 42. 

Accusative Plueal Masculine. 

This case ends in -dn. The Rik has 908 such forms, from 256 
stems. In the samhitd, however, they appear as -dn before a 
vowel, except at the end of a pdda (Rik Pr. iv.26). Thus, i.92.7, 
iisho gdagrdn tipa mdsi vd'jdn. I have noted the smhhitd-torm 
for 200 instances. In 73 it was -dn : before a-, in 42 ; before d-, 
in x.53.2; before i-, in six (as iii.32.6) ; before i-, in ix.97.56 ; 
before u-, in eighteen (as iv.22. 10); before r-, in v.32.2 and 
vi.15.18; before e-, in i. 1.2 and iv.54.5 ; before o-, in iv.4.4. 
It stood at the end of a pdda, in 68 instances unchanged ; 
but as -dn in iv.1.17; 19.7: viii.64.1 (there are six other 
exceptions given Rik Pr. iv.26, 27) ; it stood before consonants 
unchanged in 47 cases ; before p- the sandhi was -n ch- in 
i.35.5; 126.2: ii.12.10; and -n c- before c- in four cases. 
Only twice did -dnp appear, x.51.8 before ca, and v.31.2 before 
cid. See the Pr. iv.32 for the other instances. According to the 
Pr. iv.33, there are only five examples of s inserted between -dn 
and t-. Professor Whitney (note to Ath. Pr. ii.27) regards dn 
before a vowel as, equally with dns, dnp, inr, tinr, coming from 
the original ending in ns; in all these cases alike, s is treated after 
a nasal vowel as it would be after a pure vowel : e. g. jdnd dnu 
and jdndn dnu are parallel results, the one from jdnds and the other 
from jdndns. With regard to the inserted t, between -dn and s-, see 
the Ath. Pr. ii. 9. Aufrecht has left them out in the second edition 
(see Preface, p. vi.) ; but they seem to be justified phonetically ; and 
their development is aided by the false analogy of many instances in 
which the "inserted t" is organic: thus, i.80.10, indro nir ahant 
sdhasd sdhah; x.40.12, d' vdm agant sumatir; x.20.5, minvdnt 
sddma purd eti; so vii.104.21 d. 

The only instances of resolution of the d of -dn are : pukrdan in 
iv.2.2; devdan in viii.64.2; and in x. 12.2, unless devd['si]devd'n 
paribhU'r rt'ena be better; in x.93.2, finally, the metre is in a 
hopeless state. 

Nominative, Accusative, and Vocative Plueal Neuter. 

There are two endings, -d and -dni. The forms in -d number 
1682 (from 394 stems) ; those in -dni, 1050 (from 280 stems). The 
older form, therefore, which is entirely unknown in the later 
language, here greatly outnumbers the younger. The Atharvan 



N.A.V.p.n.] Noun- Inflection in the Veda. 347 

has in all 302 forms in -d (from 102 stems) against 407 in -dni 
(from 158 stems). Deducting those that occur in Rik-passages, 
the AV. has in verses peculiar to itself 228 forms in -a against 334 
in -dni. For the Rik the percentages are 62 : 38 ; for the Atharvan 
41 :59. In the Rik the older form predominates largely; in the 
Atharvan, the younger has already gained the advantage of its 
rival. Nevertheless, it seems from this, that in the AV. the 
-a-forms have held their own, as against the -<2m'-forms, better 
than any other old form, as against its more modern equivalent. 

I have not found a single example of the V.p.n. in the Rik. The 
Atharvan has one, cittdni, iii.2.4; but even here the MSS. read 
cittd'ni, according to Professor Whitney's MS. Index. 

The genesis of the younger form calls for a word. On the one 
hand, it is not a double formation (like the oft-quoted prtstishu), 
made by adding the general ending, i, to the existing form in -d, 
with mediating n; nor is it a good explanation to say that it was 
formed by adding the general ending, i, to the stem, with 
mediating n — which would give only -a-n-i (as, yugdni). We 
have here a plain case of transfer to the aw-declension. In this, 
the distinction between "strong" and "weak" stem exists, and 
the N.A.p.n. ends in -dn-i. The coexistence of such forms as 
krtrimd krtrimdni, pu"rvd pU'rvdni, yugd 1 yugd'ni, vdmd' 
vdmd'ni, talpdl talpdni (AV. xiii.1.17: xiv.2.41), shows how 
easy was the transition from krtrvma to krtriman, p&'rva to 
p4'rvan, yugd to yugdn, vdmd to vdmdn, talpd to talpdn. This 
theory is strengthened by the abundant converse facts in regard to 
the aw-stems. Moreover the form dhdndm, viii.22.13, proves 
beyond a doubt the transition from the an- to the a-declension, for 
a case other than those in question. 

The older and the younger form stand side by side so often that 
this is almost the rule. Common stock-phrases illustrate it, as 
vipvdni kd'vid. It is worth noticing that generally, if two 
nominatives or accusatives p.n. stand in the same pdda, and the 
one ends in -dni (from a- or an-stem, without distinction), -ini, 
-ilni, or even -dnsi, -inshi, -Unshi, the other eiids in -d. From 
very many examples I cite a few : ix.61.30, yd' te bhimd'ni 
d'yudhd, and vii.67,9, dpvid maghd'ni; so viii.41.5, veda nd'mdni 
guhid; i.166.10, bhu'rini bhadrd': cf. vii.19.4 b; iii.36.8, sdvand 
puru'ni; viii.67.8, vdsHni sdmgatd; iii.4.4, ilrdhvd' poci'nshi 
prdsthitd rdjdnsi; iii.1.8, rabhasd' vdpilnshi. It is plain that this 
common collocation is favored by the metre. A comparison of 
the similar metrical structure of iii.39.2c/ 55.105; vi.62.2c, and 
x.56.56, is instructive. 

The difference in the metrical value of the two forms often 
decided the poet's choice. Compare x.82.Sb, with d {bhuvandni, 
bfv&vand) ; v. 79. 7, y'e no rd'dhdnsi dpvid gavyd' bhdjanta 
stirdyah, with vi.44.12, indro rd'dhdnsi dpvidni gdvyd, and 
viii.34.14, d' no gdvydni dfvid; iv.8.3, dd'ti priyd'ni cid vdsu, 
with vii.32.15, ye dddati priyd' vdsu; x.120.7, dta inoshi harvard 
puru'ni, with AV. v.2.6, dta invata Jcdrvardni bh-d'ri; x.165.5, 



348 0. R. Lanman, [a-stems. 

samyopdyanto duritd'ni vipvd, and AV. vi.28.1, samlobhdyanto 
duritd' padd'ni. The difference between x.85.18, vipvdny anyo 
bh/dvan&bhicdshte, and vipvdnyd bhdvand vicdshte (AV. vii.81.1, 
et alibi), furnishes a most striking example of haplographia. 

In several cases, the change of -d to -dni gives a plausible 
emendation: i.173.4, td'\ni\ karma dshatard\ni\ asmai ; v.41.10, 
pocishkepo ni rndti vdnd\ni\ ; vi.22.6, dcyutd\ni~\ cid vtditd' 
suojah (these are suggested by Grassmann) ; vii.4.2, sdrh yd 
vdnd[ni~\ yuvdte p&cidan (cadence? see Kuhn, Beitrage, iii.476). 
For AV. iv.4.4, see below. For viii.12.28 30, Grassmann proposes 
bhdvand ni yemire (as x.56.5), for bhdvandni yemire. If he is 
right, ix. 86.30 needs the same change. 

In the trochaic pdda, viii.2.30 5, the resolution, ukthdd 
ca tHbhya\m\ td'ni, seems to be the best way out of the metrical 
difficulty; if correct, it gives an example of the organic form of 
the case. 

The rule of hiatus does not hold for the final -d; it sometimes 
coalesces with a following vowel : e. g., x.60.5 a; iv.25.6, 
kevalendrah; so in hdvemd', vii.29.3. This is probably the case 
in i.85.9 and viii.85.19, where the text has ndri dpdnsi, and we 
must read with Grassmann ndrid'pdnsi, i. e., ndrid dpdnsi (cf. 
verse 21). 

In several passages the pada has -d, where the assumption of 
the usage so common in the Atharvan (Pr. ii.56) is, perhaps, 
preferable: i.57.2, nimneva for nimndm iva; x.40.9, nivan'eva for 
nivandm iva; v.66.2, vrat'ma for vratdrn iva ( Ueb. i.577) ; v. 73.8, 
ydt samudrd'ti pdrshathah, for samudrdm dti. AV. iv.4.4 b, sd'ra 
rshabhd'ndm, p. sd'rd, BK. interpret as sd'ram rshabhd'ndm. If 
sd'rd may be considered a voucher for the use of the word in the 
plural at all, I should prefer to read sd'rdni rshabhd'naam. The 
metre is thus entirely rectified. For yugeva and ndbhyeva, 
ii.39.4, see N.Ad.m. 

The pada sometimes has a visarga, where the exegesis demands 
a plural neuter: thus vi.29.2, ndrids (cf. vii.45.1); i. 174.8 and 
v.29.15, ndvyds; v. 61.16, purupcandrd's. Vice versa, the visarga 
is wrongly omitted, before s-, in iii.32.5, apd drnd[s] sisarshi 
(cf. i.174.2). 

The final long -d has disappeared in Greek and Latin, leaving 
only scanty traces: II. xiii.22, acpSira aisi ; so avxa (avxoa) ; 
certd,falsd, C.I.L. i.1440, 1441 (Bttcheler, Lat. Declination, p. 19). 
On the other hand, in the Veda, there is hardly a trace of this 
shortening (save in forms from cm-stems), although, as syllaba 
anceps, it might often be read as short. There is reason, however, 
for assuming this shortening in some cases: vdna, ix.88.2: 
stiarshdtd vdna tirdhvd' navanta, p. vdne (cf. i.88.3, and 171.3, 
and Grassmann's note) ; dha, i.92.3 : vipved dha ydjamdndya 
mnvat'e (cf. i.130.2, dhd vipvd) ; possibly vipva[m] in vn.7.2, 
jdmbhebhir vipvam upddhag vdndni (but see Grassmann's note), 
with -m as in i.60.3, jd'yamdnd (text, -nam as-), which is certainly 
a N.s.f. Cf. patdmilti, satydmugra, sahdsram-dti; and ehamanta, 
Indische Studien, v.437. 



N.A.V.p.n.] Noun- Inflection in the Veda. 349 

The difficult verse, ix.10.9, abhi priyd' divas paddm adhvar- 
ydbhir guhd hitdm (sc. hav'er dpatyam) su'rah pagyati cdkshasd, 
offers perhaps the least uncertain example of this phenomenon. I 
translate : ' To its [own places of the sky] home in the sky, the 
Soma, set away [in the vat] by the priests, looks with the eye of 
the sun ;' i. e., with the image of the sun reflected in its surface, it 
seems to look aloft to the regions whence it was brought by the 
falcon (iv.26.6). How priydm could become priyd' (see Grass- 
mann) in our text, I cannot conceive ; but that paddm can here 
stand for padd' is even probable; cf. ix. 12.8. Of course, the 
Sama variant smooths everything over and has priydm (ii.477) ; 
but its reading ought not to have any weight here. 

Gr. says that vdpa is used once as neuter, in ii.24.13 (p. vdpd). 
By BR. it is described, s.v., as standing for vdgdn. It is a simple 
case of elision (before r-, for vdpam), and crasis; pronounce 
vdparndm. So Roth would now correct the dictionary. Veshdnd, 
iv.33.2 (Gr., A.p.n.), is probably an I.s.n. 

Some of the most frequent forms are: in -a, imd', 57; uhthd! , 
25; td', 71; duritd', 31; bhuvand, 36; yd', 50; vipvd, 249; 
sdvand, 30 ; havyd', 44 : — in -dni, krtd'ni, 20 ; td'ni, 34 ; 
bhtivandni, 57 ; yd'ni, 25 ; vipvdni, 110; vrtrd'ni, 36 ; vratd'ni, 
34 ; havyd'ni, 25. 

Instrumental Plural Masculine anb Neuter. 

There are two endings, -ebhis and -ais. In the Rik, instru- 
mental in -ebhis are almost as frequent as those in -ais, and in 
certain parts more so (details will be given hereafter). The forms 
in -ebhis number 571 (414 from 146 m. stems, and 129 from 65 n. 
stems, besides 28 occurrences of the pronoun tebhis) ; the forms in 
-ais, 666 (457 from 142 m. stems, and 209 from 79 n. stems). The 
Atharvan has in all 53 forms in -ebhis against 263 in -ais; and of 
these, in passages peculiar to itself, it has only 43 against 226. 
The ratio for the Rik is nearly 1:1; for the Atharvan, about 1:5; 
that is, the ending -ebhis is plainly going out of use in the Vedic 
period. The later Sanskrit uses the old ending only in one word, 
the pronoun ebhis; and it has been retained here doubtless because 
the word would otherwise have lost its individuality almost 
entirely. The two forms often stand side by side, as in i.33.2, 
upamebhir arkaih; A V. vii. 98. 1. Once, in place of the Rik-reading 
-ebhis (x.14.5, dngirobhir a' gahi yajfayebhih), the Atharvan has 
-ais (xviii. 1.59, dngirobhir yajmyair d! gahi'hd). The choice was 
often determined by the needs of the metre. Compare ddityebhir 
vdsubhir dngirobhih (vii.44.4), and ddityai rudrair vdsnbhih 
sacdbh'dvd (viii.35.1); ydtdm dpvebhir agvind (viii.5.7), and 
ddityair ydtam qpvind (viii.35.13). 

Bopp ( Vgl. Gr. 3 §219) considers both -ebhis and -ais as two 
independent developments from -dbhis. But besides the two 
forms cited by him, asmd'bhis and yushmd'bhis, I know of no other 
to support this view. Neither of the two seems derivable from 
the other. The equality of their numbers, at least does not run 
vol. x. 48 



850 0. R. Lanman, [a-stems. 

counter to Bopp's view. The resolution of -ais into two syllables 
would speak for this theory ; but I cannot bring forward a single 
certain example of such resolution. It has been proposed in the 
following passages: at the end of octosyllabic pddas, viii.92.13 5 
and x.24.2 a; but in view of the facts relating to the resolution of 
devds (N.V.p.m.) etc., it may be asked whether even such pddas 
were not sometimes catalectic. In i.129.8 and v.50.2, the metre 
can be helped by better means than the resolutions atrdis, 
sacathydis (Kuhn, Meitrdge, iv.192) : read upa-ishe (tr forms weak 
position) and sacathiais. In the trishtubh line ii. 11.2, read dva 
abhinad, not ukthdis. In v. 4. 7, the syllable before ukthais is 
syncopated. In viii.90.10, rdjishthdis is not needed, since pddas 
of eleven syllables are sometimes combined with those of eight. 
Cf. also Kuhn, Beitrdge, iv.189. 

Specimen forms are : in -ebhis, ukthebhis, 14 ; ebhis and ebhis, 
20; t'ebhis, 28; devebhis, 52; yebhis (yais not found in RV.), 28; 
vd'jebhis, 21; vipvebhis, 29; stomebhis, 26: — in -ais, arkais, 43 ; 
dpvais, 30; ukthais, 35; devais, 31 ; yajndis, 34; stomais, 25. 

Dative Plural Masculine and Neuter. 

The dative plural ends in -ebhyas (or -ebhias) and this form does 
duty also for the ablative. The datives number 131 ; that is, 79 
masculines in -ebhyas (from 28 stems), and 51 masculines in -ebhias 
(from 26 stems), and one neuter. 

1. The masculines are as follows: arbhakebhyas, ddit'ebhyas, 
dpinebhyas, ebhyas 8, ebhyas 2, ubhdyebhyas, ekebhyas, etebhyas, 
grh'ebhyas, jivebhyas, t'ebhyas 4, tdvakebhyas, ddpayoktrebhyas, 
devibhyas 36, pipunebhyas, putr'ebhyas, pHrushebhyas, pU'rvebhyas 
4, mdrtiebhyas, md! nebhyas, yebhyas 3, rdthebhyas, viprebhyas, 
vipvebhyas, viapvebhyas, suvidatr'ty ebhyas, stenebhyas, pasprdhdn- 
ibhyas. 

Of the neuters my collections show only one single example 
in the whole Rigveda, and that is bMtvanebhyas, viii.85.16. 
It might be supposed from the grammars that the forms of the 
paradigms were of about equal use and frequency. But when we 
compare the occurrences of the N.s.m. (10,071) with those of the 
D.p.n. (1), the contrast is striking. See Delbruck, Verburn, pp. 14, 
15 ; and Curtius, Verbum, vol. i., preface, p. v., and p. 5. 

2. The ending of the D.p.m.n. is frequently to be pronounced 
as a dissyllable. The instances number 51. 18 are at the end of 
a jagati-pdda whose cadence would be catalectic without the 
resolution: ajdrebhias, x.94.7 ; drnavattarebhias and dpdapastar- 
ebhias, x.76.5; ebhias, i. 146.5 : iii.53.16; jdnebhias, i.55.5 ; 58.6; 
tebhias, iii.2.6; ddpakakshiebhias and ddpayojanebhias, x.94.7; 
pdrvatebhias, iv.54.5 ; pitukr'ttar ebhias, x.76.5 ; mdrtiebhias, 
x.17.2; tndnavebhias, iv.54.1 ; md'nushebhias, iv.54.2 ; yajniy- 
ebhias, i. 139.7: iv.54.2; s6marabhastarebhia$,x.T6.5. Further, in 
pddas of 11 or 12 syllables, -bhias occupies the third and fourth 
places seven times: gdrbhebhias, i. 146.5; devebhias, i. 139.7: 
iii.34.7 : x.70.2; 110.4, 5; putrebhias, x.15.7; — the sixth and 



D.p.m.n.] Noun- Inflection in the Veda, 351 

seventh in x.32.5, ti'mebhias, and i. 122.8, pajrebhias; — the ninth 
and tenth in v.11.1, bharatebhias, and iv.41.3, papamdnebhias. 
At the end of octosyllabic pddas, -bhias occurs 1 1 times : ebhias, 
i. 131.5 ; jdnebhias, x.156.4; dev'ebhias, x.85.17 ; bddhit'ebhias, 
iv.30.4 ; mdrtiebhias, i.90.3: viii.9.16; yajatebhias, ii.5.8; yajmy- 
ebhias, i.139.7: v.52.5; vid'dshtar ebhias, viii.41.1 ; somiebhias, 
viii.32.5 ; — and occupies the fifth and sixth places of such pddas 
in 10 instances: dev'ebhias, i. 13.11 ; 142.6; 188.10: v.5.11: ix.3.9; 
28.2; 62.20; 99.7; 103.6; putr'ebhias, vii.32.26 ; finally, in the 
dvipadd virdj, ix.109.21, we read dev'ebhias tvd. 

Ablative Plubal Masculine and Neuter. 

My collections show a total of only 36 ablatives plural 
(masculines 25 and neuters 11). In view of this fact, it is easy to 
understand how a relation of so infrequent occurrence should have 
no special form of its own, but depend upon some other case-form 
for its expression, as here upon the dative. 

1 . The masculines in -ebhyas are : ugrebhyas, ix.66. 1 7 ; grh'ebhyas, 
i.120.8; devebhyas, i.74.9: iii.9.5: vi.44.7 : viii.90.16 :' x.97.1 ; 
pdrvatebhyas, x.68.3; makh'ebhyas, vi.66.9 ; pU'rebhyas, ix.66.17. 
(10). The neuters are : dntrebhyas, x. 163.3 ; padebhyas, viii.2.39 ; 
vdnebhyas, ii. 1 . 1 ; vipvebhyas, ii. 23.17 ; harmi'ebhyas, vii.76.2. 

2. Vocalization of the written y takes place in 21 instances, as 
follows : at the end of octosyllabic pddas, in dsurebhias, viii.86. t ; 
ebhias, vi.46.9 ; jdnebhias, i.7.10 ; pdrvatebhias, i. 191.9 : viii.34.13 
(allm.) ; and in the following neuters: (dveshobhyo) , nydkrtebhiah, 
viii.68.3 b, which we must read with elision and vocalization ; 
durit'ebhias, viii.44.30; nakhebhias, x. 163.5; pd'rthivebhias, 
x.158.1 ; — ■ -bhias occupies the fifth and sixth places of octosyllabic 
pddas in dntebhias, i.49.3 : viii.77.5 ; jivebhias, viii.8.23 ; dev'ebhias, 
ix.42.2 ; 65.2, 3; viprebhias, x.135.4 (all masculine); — and in one 
neuter, nirdhrebhias, viii.44.3. At the end of a jagatt-pdda stand 
the masculines ebhias, x.64.2, and pdrvatebhias, vii.104.4; in 
bhtivanebhias (n.), ii.23.17, ia fills the ninth and tenth places: — 
pyenebhias (m.), finally, is in the middle of a trishtubh-pdda, 
iv.26.4. 

Genitive Plural Masculine and Neuter. 

This case ends in -dndin. The oxytone stems in d never shift 
the accent to the endings as do those in i, ■&, r' (except in the 
numerals). The series of development, so far as illustrated by 
actual Vedic forms, would be as follows : «. devd'm (for devddm), 
b. devdndm, c. devd'ndm, d. devd'naam. 

a. The forms like devd'm, as the organic ones, are d priori 
probable, and are supported by the analogy of the Zend. So 
a$tdm, Yd. vi.29; xv.3, with* ahmarstanam; geredham Vd. 
iii.10, 22 ; ptaordm Vd. viii.12, etc. See Justi, Handbuch, § 528.4. 
Geldner (Metrik des jilngeren Avesta, § 70) reads — as I think, 
rightly— daev dm, Yc. x.6. But I can discover no such form in the 
Gathas. Were it not for this probability, however, these forms, 
of which the comparative philologists make so much, would not 
seem very well vouched by the Veda. They are given under 3. 



352 C. R. Lawman, fa-stems. 

b. Fovms in -andm with short penultima are written nowhere. 
A careful search of all the -dndm-iovms, with scansion of the pddas 
in which they occur, revealed only five instances in which there 
was even metrical evidence of a short thematic -a: i.44.2 
(== viii.11.2), dgne rathi'r adhvard'ndm; i. 188.11, purogd' agnir 
devd'ndm; x. 1 3Q.fi, apsardsdvi gandharvd'ndm; x. 1 74.5, ydthdhdm 
eshdm bhUtd'ndm. viii.70.7 b is trochaic. A short is merely 
favored in the seventh place of i. 124.1 1 6 .' vii. 5.1c- x.2.6a, and 
elsewhere. These five instances do not weigh much, and evidently 
belong to later parts of the Rigveda. But this short a is well 
authenticated in Zend, as Yt. x.65, yd aredranam aredrd (Skt. yd 
radhrd'ndrh radhrdh) ; so takhmanam, vydkhnanam, ibid. ; 
Yt. v.21, appandm arshnam (Skt. dpvdndm vr'shndni), etc. 

c. The G.p. in -dndm occurs 370 times. The masculines number 
333 (from 59 stems) ; the neuters, 37 (from 17 stems). That the 
long d before n is a lengthening, the forms in -indm, -tindm, -rndm 
seem to indicate. It is explained by Job. Schmidt ( Vocalismus, 
i.39) as due to the following nasal. Osthoff (Forschungen, ii.17, 
and in Paul and Braune's Beitrdge, iii.3 and 39) explains the n as 
due to a transfer from the a- to the an- declension. 

d. Kuhn (Beitrdge, iv.180) notices the resolution of the d of the 
G.p., as in devd'naam. This seems to me to be a purely metrical 
phenomenon, and to shed no light on the genesis of the form. 
Bezzenberger takes an entirely different view of it (in his 
Beitrdge, 1877, ii. 130 ff.). Forms of this kind occur 157 times. 
Of these 144 are m. (from 49 stems) ; and 13 are n. (from 7 stems). 
They are distributed as follows : mandala i. has 32 ; iii., 3 ; iv., 8 ; 
v., 15; vi., 9; vii., 5 ; viii., 59 ; ix., 9; x., 14 ; Val., 3. The 
resolutions at the end of vi.48.12e: vii. 32.11c: viii.19.33c; 23.2 6; 
90.6c: x.23.\b; 103.4<?, are such as convert the cadence from 
trishtubh to jagati. A very large majority of the resolutions ( 1 06, 
or two third's) are at the end of an octosyllabic pdda (and 40 of 
the 106 are in the eighth mandala). This fact suggests again the 
question whether such pddas may not often be regarded as 
catalectic. There are examples enough of the resolution within 
the pdda, however, to put the validity of this metrical process 
beyond question. 

1. Examples of this case are : adhvard'ndm, 1 ; ddityd'ndm,\\ ; 
jdndndm, 34; devd'ndm, 148; yajniydndm, 12: — dhdndndm, 13. 

2. The cases which suffer resolution are : 

Masculines: adhvard'naam i.1.8; 44.9; 45.4: iii.10.4: iv.7.3 : viii.8.18. amr'td- 
naam x.33.8; 14.3. ard'naam viii.20.14. afvdnaam v.18.5: vi.63.10. apoidnaam 
viii.25.23. dmtdnaam viii.53.3. dditid'naam viii.18.2. ■ukshitdnaam v.56.5. 
v,mhalasv,tdnaam i.28.1, 2, 3, 4. ushtrdnaam viii.5.31. kalafdnaam iv.32.19. 
Jcr'tvidnaam viii.25.23. garddbhd'naam Vii. 8.3. jandnaam with T.,vi. 45.9. jdnd- 
naami.25.li: 75.3,4; 81.9; 191.4: iv.9.5 : v.16.2 ; 66.4 : vii.16.2, 7 ; 56.24; 
74 6: viii.1.4; 5.13; 15.10; 19.33; 24.4; 45.28; 53.3; 92.6: ix.52.4 ; 64.27; 
112.1- Val. 6.7. turd''aaam vi.48.12. devd'naam i.43.5 ; 50.5; 133.7; 187.6: vi.15. 
13- viii.31.15; 39.6; 58.3; 64.8; 83.8: x.93.3. navdnaam viii.5.37. nemdnaam 
vi.16.18. parvatdnaami.39.3; 187.7: v.84.1: viii.18.16; 31.10. poshi&naam iv.48.5. 
priyd'naamvui.92A0. bhardnaamY'mAO.3. maddnaamviii.S2.3\: ix.104.5. madd- 
naam iv.31.2 : viii.81.6: ix.23.7. mdrtdnaami.63.5: iv.2.18: viii.39.6. mdrtidnaam 



G.p.m.n.] Noun- Inflection in the Veda. 353 

i.26.9 : v.9.6 ; 74.7 : viii.67.6 ; 90.6 : x.33.8. mahd'naam i.187.6 : viii.52.1 ; 81.3 ; 
83.8: ix.1.4. md'nusMnaam i.84.2 ; 127.8; 128.7: iv.8.8; 9.5: v.7.3: viii. 23.25 ; 
41.1 : ix.61.11. yajnd'naam vi 16.1 : x.26.5. y&'dudnaam viii.6.46. rdthdrmam 
i.48.3 : v.52.9 ; 53.10; 66.3; 74.8; 86.4 : vii 32.11 : viii.23.2 ; 57.4; 83.1; 92.10: 
x.26.5; 78.4; 103.4. rdsdnaam i.187.5. rudriydnaam viii.20.3. vatsd'naam 
vi.24.4. vd'jdnaam viii.24.18; 81.3 : ix.31.2 : x.26.7, 9. vdjdnoam i.29.2 : vi 45.10 : 
viii.81.30. viprdnaam x.26.4. vivratdnaam x.23.1. vr'kdnaam viii.56.14. vrtra- 
hdthdnaam iii.16.1. vrshabhd'naam Val. 5.1. (aiisidnaam i.17.5. fdkund'naam 
ix.112.2. faphd'naam v.6.7. f&'rdnaam vi.68.2. sutd'naam viii.32.19; 34.5; 
53.3. somdnaam i. 134.6: viii.82.33. somid'naam viii.17.14. svddhainavdnaam 
viii.32.20. havdnaam viii.26.16. hvdrid'naam v.9.4: — neuters, aghd'naam viii. 
47.2. ahdnaam viii.22.13. diydnaam viii. 19.37. bhuvandnaam viii.41.5. rddhd- 
naarn i.30.5: iii.51.10. vd'ridnaam i.5.2 ; 24.3: viii.60.11, 13: x.9.5 ; 24.3. 
sajdtidnaam viii. 7 2. 7. 

3. The instances of a G.p. in -dm are not all entirely beyond 
question. They are as follows: yrtthidm in Val. 8.4, yUthidm 
dpvdndm, ' of horses belonging to the herd :' here there is no 
doubt about the form ; but the hymn in which it occurs is very 
poor stuff. Cardthdm in i.70.3, gdrbhap ca sthdtd'm gdrbhap 
cardthdm, ' and child of the things that move not (wood, flint) — 
child of them that move (clouds, lightning) :' Say. caranavatdm 
jangamdndm garbhah. Hinsdndm in x.142.1, are hinsdndm dpa 
didytirn dl krdhi; BR. take it as G.p.m. of hinsa, but I can find 
no occurrence of this word elsewhere in the Veda. Grassmann 
( Wb. 1665) makes it a participle (cf. v.64.3, dhinsdnasya, and for 
the accent, viii.43. 10, ninsdnam), 'das verwundende Geschoss;' 
but as didyti, is not feminine, it is quite as easy to consider 
hinsdndm as standing for hihsdndndm: 'Of those that would 
harm us, turn far away the dart.' Here the metre supports the 
shorter form ; but a real instance of haplographia is girvdhas, 
vi.24.6, if Gr. reads aright, with SV. i.68, girvavd'has. So too, 
perhaps the easiest way to dispose of the inferior reading of SV. 
i.340, tirdh purOf cid arnavd'm jagamyd'h (RV. x.10.1, arnavdm), 
is to make it a G.p.m. : ' E'en if o'er many a billowy sea thou 
fleddest.' In x.46.5, ndyanto gdrbham vand'm dhiyarh dhxih, 
' Solemnly they led the offspring of the sticks of attrition,' vand'm 
may stand for vand'ndm (accented as in iii.9.2). The stem van is 
supported only by vdnsu, found twice, with irregular accent. 
SV. i.74 has a wretched variant, vand '. I regard pdsd'm as G.p.m. 
of pdsd, 'ruler,' in ii.23.12, pdsd'm ugrd mdnyamdnah, 'deeming 
himself the mighty one [of the] among the rulers.' See BR. 
vii.168, and Gr., Ueb. 1.570. 

In the six cases above, the text has a final m; in the following 
six it has n or w, for which we must read m or m. In vi.47.16, 
ooshkHydte vipa indro manushyCCn, ' The tribes of men Indra takes 
to his protection,' I consider manushySn (read -d*m) as equivalent 
to manushyd^ndm. The only alternative, to regard it as in 
explanatory apposition to vipas, is a bad one. In iv.2.3, antdr 
lyase arushd! yi/jdnd yushmd'np ca devd'n vipa d' ca mdrtdn, we 
must read mdrtdrn, and interpret it as G.p.m. : ' Thou goest as a 
messenger — .... to you the gods, and hither to the dwellings of 
(us) mortals.' Sayana felt the difficulty, and says shashthyarthe 



354 G. R. Lawman, [a-stems. 

dvitiyd. In iv.2.11, cUtim dcittirh cinavad vi vidvd'n .... mdrtan, 
' The wisdom and folly of mortals let the wise one distinguish,' 
read mdrtdm as before. TS. v.5.4 4 is without variant. In i.71.3 
and vi.11.3 we have devd'n jdnma (p. devd'n); and in x.64.14, 
devd'n jdnmand (p. devd'n). Here we must read in the Samhita 
devd'm, and in the Pada devd'm, and interpret, with Grassmann, 
as G.p. The metre requires a change from devd'ndrh to devd'm 
in vi.51.2. Finally, Bollensen's simple transposition of dev6 and 
devd'fidm sets the metre right in i.68.2. 

Locative Plural Masculine and Neuter. 

This case ends in -eshu, and its final vowel is regularly uncom- 
bined. Of such forms there are 548 masculines (from 123 stems), 
and 249 neuters (from 92 stems), or 797 in all. In some cases the 
gender is doubtful, as vd'resku. In all these 797 instances, the final 
-u, as shown by actual scansion, 1. is never changed to -v before a 
dissimilar vowel (one exception) ; and 2. it never coalesces with a 
following similar vowel. (Cf. Benfey, Gram. p. 51, § 86, end.) 
When standing, as often (so adhvar'eshu, mdrtieshu), at the end 
of a pdda, it is written -v if the next begins with a vowel. Here, 
as is well known, it is always pronounced with hiatus (so i.27.5 : 
vi.49.2; 64.4: viii.85.19: x.30.8; 95.16; 112.4: i.154.2: iii.23.1 ; 
34.3: viii.9.10; 16.5: x.103.11; before w-, vi.4.2). Aside from 
these, there are 179 cases (m. 125 ; n. 54) where -u is followed by 
a vowel in the same pdda, and it is invariably to be read with 
hiatus: thus, before «-, i. 134.5: iii.57.5 : iv. 18.4; before the 
postposition dl , 44 times, as ix.8.6 ; before o-, i.59.3 ; 91.4 ; 108.11 ; 
before e-, i.64.4 ; 166.10: ix.102.3: x.28.2; before r-, iii.10.2 : 
v.44.5: vi.59.4: x.21.7; before «-. iv.7.1: v.14.2; 22.1: viii.11.1 ; 
49.3: x.21.6; before i-, 16 times, ii.11.36, 17: iv.32.11: v.28.4 : 
viii.81.26; 88.2: x.63.9: ii.ll.3a: iii.41.4: iv.21.4; 22.5; 30.16: 
vi.46.1: vii.99.6: ix.97.56 : x.42.4 ; before a-, 24 times, as v. 25Ab; 
and before a-, (51-f-24=)75 times. In i.l62.l(5d, I propose the 
reading abhi for d! ; in \.f>\.hd, we must of course pronounce 
prdryi$vdnam. The solitary exception to the rule of hiatus is 
x.121.8, yd deveshv ddhi devd eka d'sit. This hymn is confessedly 
a late one, and the sandhi an additional proof of its lateness. It 
is quite possible to read vi. 26.2c metrically as it stands: tvd'rh 
vrtreshv indra sdtpatim tdrutram; but our canon shows that we 
must pronounce (tud'm) vrtreshu, and the word indra must 
accordingly be thrown out as a gloss. Herein I find that 
Grassmann agrees with me. 

The Veda, then, does not support the view that -sva was the 
original ending of the locative. Its combination with the post- 
position d', which, if any thing, ought to serve as the point of 
departure for this theory, is found only 44 times in 797, and never 
as svd. Cf. Bopp, Vgl. Gr. 3 § 279 note. 

Examples of the forms are : masculines, adhvar'eshu, 27 ; deve- 
shu, 99; mdrtieshu, 25 ; yajfieshu, 35 ; vd'jeshu, 41 ; sut'eshu, 16: — 
neuters, uhth'eshu, 10 ; bhUvaneshu, 12 ; vdneshu, 20 ; viddtheshu, 
33 ; sdvaneshu, 14. 



N.s.£] Noun-Inflection in the Veda. 355 



STEMS IN A. 

The following is a synopsis of the terminations of the 
inflectional forms : 

Singular: N., -d, -aaf, -a?; A., -dm, -aam, -d? ; Y.,-d,-ayd; 
D., -dyai, -ai; Ab., -dyds, -dyds?; G., -dyds, -dyds?, -dyaif '; L., 
-dydm; V., -e. 

Dual: N.A.V., -e (iti); I.D.Ab., -dbhydia; G.L., -ayos. 

Plural: N.V., -dsas, -ds, -<?'/ A., -ds, -aas,-dsas?; I., -dbhis; 
D.Ab., -dbhyas, -dbhias; G., -dndm, -dnaam, -dm?; L., -dsu. 

There are no truly declensional peculiarities of accent. 

Nominative Singiulab Feminine. 

The case-form is regularly like the stem, and occurs 1053 times 
(from 424 stems). Examples are: idd, 17; eshd',27; citrd',9; 
jjdydl, 17 ; ddkshind, 21 ; yd', 74 ; yoshd, 24 ; sd', 80 ; subhdgd, 13 ; 
stinr'td, 10. 

Bopp (Ygl. G-r. 3 §137) and Schleicher (GompA §246) assume 
for all these stems a N.-ending s, which has become lost. This 
view is controverted by Max Mtlller, Chips, iv. p. 46 ; and 
Aufrecht finally (Migveda, 2 preface, p. v, note) calls it a " dream 
of the comparative philologists." The Veda does not support it. 
The s of gnd's, iv.9.4, proves nothing. Whether it be from Jaw or 
jfid, it was originally dissyllabic, but is treated as a root-word. 
So stri' has come to be treated declensionally as a root-word, 
preserving a trace of its true character {sUtri) only in the N.s., 
where the s is lacking. 

The final d often stands before an initial vowel in the text. Its 
treatment by the samhitd-pdtha is varied. In about 30 instances 
it stands at the end of pdda a or c, and is always written with 
sandhi; but it is of course always to be pronounced with hiatus : 
as, asa-d ca yd' na urvdrd : d'd . . . , viii.80.6. The fusion results 
oftenest in d, as i.H4.9c/ 123.11a: ii.27.7a; iii.39.la/ 58.1«; 
sometimes in e, as i.32.9a; iii.-S9.lcy or o, as iii.58.lc v.80.2c (cf. 
3a), 5a: vi.61.13e/ 64.2c, 5a; vii.95.4a: viii.20.4a. The result is 
a at the end of a pentasyllable pdda, i.66.7a and iv.10.5a. Fusion 
resulting in ai or au is not found (Rik Pr. ii.31). Accordingly we 
have, i.123.10, pd'gaddndn : eshi. In v.30.14, we have yd'n: rnath- 
caye, because the individuality of the word (if written yd) would 
be obscured. Of the 74 occurrences ofya'as N.s.f., only two are 
before vowels : this is the only one to which Rik Pr. ii.31 (end) 
applies; the other is vi.64.5. A merely graphic peculiarity is 
the short a in priyd:r'-, i.151.4, and rjuhdsta:r-, v.41.15. The 
pada of course has d. 

In 160 instances the final d stands before a vowel in the interior 
of a pdda. It is always written with fusion, but the scansion 
shows that the real treatment is determined by the metre. Thus 
i.95.1,anyd'nyd vatsdm tipa dhdpayete; but i.62.8, vdpurbhir d' 



356 C. R. Lanman, [^-sterns. 

earato any d'- any d; cf. iv.52.2, dpveva citrd' drushi. In the great 
majority of these instances (137), the concurrent vowels are to be 
read as they are written : thus, vi.59.6, pd'rvd'gdt; iii.39.2, seydm; 
vii.3.9, ptiteva, and so with iva 59 times out of 61 (exceptions are 
v.45.2 and viii. 90.13) ; \Y.2.\0,prited h6t(a)rd; i.56.4, tvd'vrdhotdye 
(so eshd i.46.1, and s6 i.191.11: x.11.3; 23.4; 63.16); x.62.11, 
ydtamdnaitu. 

The final d of the dual, it will be remembered (p. 340), always 
coalesces with a following vowel. Here, however, the metre 
shows hiatus in 23 instances, as vi.64.5, sd' d' vaha yd! ukshdbhir 
dvdtd: ■&-. The list is: i.48.7c (12 syll.) ; 62.8^; 77.1a, lb; 104. 
5-a: iv.52.2a: v.45.2, sthtX'nd iva; 53.9a; 61. ha: vi.25.1, 
madhyamd' ind(a)ra; 64.5a, bis; 66.3c?.- vii.34.1, prd pukrd' etu 
(aksharapankti): viii.89.11e; 90.13, citrd' iva: ix.86.446: x.61.18c; 
93.135; 95.16a; 121.2c; 145.3a; also i.62.8e, where a syllable is 
syncopated between aktd' and ushd'h. In i.l20.1a the metre is 
hopeless. 

Besides the above, there are a number of cases given by the Rik 
Pr. ii.29, in which hiatus occurs, not only in the spoken, but also 
in the written text. It is hard to see why the diaskeuasts and 
Caunaka have taken account of these and not of the others. Those 
that concern us are: Uhd', viii.5.29 ; manishd' (as N.s.), i.101.7 : 
v.11.5: vii.70.7: vii.34.16 (at the end of an aksharapankti). At 
the end of common pddas the fusion is written, even with manishd', 
iii.33.5c: vi.67.2a. 

This comparatively frequent toleration of hiatus suggests to me a remedy for 
several lame verses: i.173.8, vipvd te anu jdshid [a]bJiM gaiih (cf. iv. 43.4a, where 
I read [a]bhud) ; vii.58.6, pra sd' [a]vdci sushtutir maghmdm (cf. vii.70.1). On the 
other hand, since fusion is common, samAiM-manuscripts would show no difference 
between saramd vidad gd'k and -md avid-, v.45.8; cf. vi.61.10, stdmid IMt; 12, 
hdvid bMt ; v.44.8c and x.99.6d And so I think that such combinations may 
have been a point of departure for the loss of the augment. See Delbriick, 
Verbwn, p. 80 ; and Curtius, Verbum, i.132, and especially 133. 

In jd'yamdnam (asmdt sukirtih), i.60.3, there seems to be a clear 
case of' a N.s.f. with shortened final and paragogic m before a 
vowel (cf. a-stems, p. 349, top). For maM, v.41.15: x.77.4, 
Gr. (Kuhn's Zeitschrift, xvi.170) reads the more organic form 
mahid. I believe he is right, and would add vi.66.3, vide h% 
mdtd' maho (sc. gdrbhasya) mahid sd'. The metre seems to 
require resolution (?) in vii.75.4, yujdndd, and x.162.1, 2, dmfvaa. 

The pada reads d in several cases wrongly : for asurid, vii.96.1, 
and patatamd'. vii.19.5, see A.s.f. ; for hantnak'eva, iv.32.23, see 
N.d.f. ; for ydshd, ix.96.24, see N.p.f. ; for sumedhd' (Gr.), m.57.5, 
see as-stems, N.s.f. _ _ 

In xA9A, priyd' may be A.p.n. ; and in 1.122.1, ishudhia , l.s.t. 
For krdnd', i. 139.1 and v.7.8, see p. 334. 

Accusative Singular Feminine. 

This case terminates in -dm, and occurs 522 times (from 226 
stems). Examples are : dmivdm, 9 ; manishd 'm, 21 ; jdyd'm, 11 ; 
yoshdm, 7 ; ydshandm, 6 ; td'm, 29 ; imd'm, 52. 



A.s.f.] Noun-Inflection in the Veda. 357 

Resolution seems unavoidable in three trishtubh-pddas : vii.93.3, 
kd'shthaam, and i.173.2, mandam (aa occupies the sixth and 
seventh places in both) ; and i.165.15, vayd'am (end) ; but see as- 
stems, A.s.m.. Gr. proposes kkaam, viii.61.7or, and mahtydmd- 
naam, iv.30.9i, thus making them full amishtubh-pddas. They 
may be catalectic. In x.50.5, Gr. reads mahidm (cf. N.s.f.) for 
mahi'm. I prefer dvamdtrdm : text, 6m-. In ix.12.7, he follows 
the Saman (ii.552), and reads sabardughdm for -ah. Bollensen 
{Orient und Occ, ii.460) supposes a copyist's error in vii.96.1, the 
loss of anusvdra before a nasal, and amends, with good reason, 
thus: asurid\m\ nadi'naam. For sumedhd'm, viii.5.6, icshd'm 
and ushd's, see cis-stems, A.s. and p.f. The A.s.f. of stems in id 
often coincides with the L.s.f. of those in t Thus pUrvid'm may 
be referred to pHrvid or to puru; satyd'm, to satyd or sat. The 
accent would distinguish pamid'm (stem pamid'; but cf. pdmids 
iii.33.13) from pdmidm (stem pdmi). Since stems in ia form their 
feminines in id or i, it is well to keep this double possibility in 
view (cf. d'ria, daivia, svaria). 

A striking example of elision and crasis is nivepane patatamd'- 
viveshih, vii.19.5, for -md'm av- (p. -mdl). Cf. Sayana : patatamim 
puram vydpnoh. 

Instrumental Singular Feminine. 

This case is formed in two ways : either the general ending -d 
is affixed directly to the stem, and the concurrent vowels 
contracted (as jihvd'-d, jihvd') ; or, a y is interposed, and the 
thematic vowel shortened (asjihvdyd). 

The older form occurs 300 times in the RV. (from 95 stems) ; the 
younger, in -ayd, 3 5 6 times (from 11 3 stems). The older form appears 
only sporadically or in stereotyped words in the later language. 

It happens comparatively seldom that the stem displays both 
endings (e. g., gavyd', dhd 1 rd, manishd') ; the reverse was true of 
-dsas and -ds. The older formation prevails especially with stems 
(generally abstracts) in td (20 out of 95), and in id or yd (44 out 
of 95). With the latter, the younger formation would sound 
badly: thus, hiranyaydyd (not found) ; cf. hiranyayd', vii.66.8. 
Here, as elsewhere, the choice between the longer and the shorter 
form was often decided by the exigencies of the metre, or by 
euphony, or both; cf. ix.98.3, dhd'rd yd drdhvd adhvar'e, with 
J01.2, yd dhd'rayd pdvakdyd; and i.91.1a, with 94.15. The 
longer form fits often in the anushtubh or jagati cadence ; the 
shorter in the trishtubh. In viii!6.33, the measure requires, 
perhaps, utd brahmanydyd vaydm: text, brahmanyd '. We find 
the final d coalescing with an initial, e. g., in ix.97.45a, and 
making hiatus in i.53.85. 

There are no purely declensional irregularities of accent. In 
the case of ayd' (from d), it is something peculiar to that pronoun, 
and not adverbial (see below, and cf. kdyd, as pronoun 11 times, as 
adv. 3 ; tdyd, 10 ; ydyd, 23 ; svdyd, 3). In aryayd', v. 75.7, BR. 
1.447, see an error for aryd[s] d', stem art 
vol. x. 49 



358 G R. Lanman, [^-sterns. 

The adverbial character of adatrayd' (from adatrd), v.49.3, 
explains the displacement of the accent ; so naktayd', iv.11.1, and 
svapnayd', AV. v. 1.8; and perhaps rtayd', RV. ii.11.12; cf. 
sarvahrdd' . The adverbial shifting of accent is natural in the 
homophonous instrumentals, since it differentiates them from the 
nominatives of the same form ; thus ubhayd' (stem ubhdy a), madhyd 
(mddhya), samand' (samand) : for dakshind' (ddkshina), cf. 
i.100.9, dakshind, oxytone, and see Indische Studien, iv.160. 

I cannot account for the accent of pavayd' in ix.97.53, utd na 
end' pavayd! pavasva, and therefore I read pavd'yd 1 , i. e., pavd' 
ayd! (cf. 52, ayd' pavd' pavasvaind' vdsuni). 

For ddkshinayd rdthena, i.123.5, see G.s.f. 

1. The homophonous instrumentals are as follows: (agriyd' iv.34.3?) abdayd', 
abrahmdtd, arhdnd 3, avi'ratd, acvayd' 2, irasyd', ishitatvdtd, ishudhid't, irmd 1 5, 
upa/md' 2, ubha/yd', wushyd', ufdnd 6, rj&yd', rtayd', etd' (x.95.2 ; Eoth would like 
to take it as A.p.n., adverbially, 'here ;' I think it is used with slightly contemp- 
tuous tone; so Say., scholion to (Jat. Br. xi.5.1', etd etayd tvadtyayd arthapHnyayd 
vdcd), aidhd', kavydtd, gavyd' viii.46.10: ix.64.4 (vii.18.7?), guild 53, (ghrnd', see 
p. 335) jardnd 2, jigishd' 2, jihvd' 3, tanyatd' 3, tdtyd', tirafcdtd 2, tejishthd, tyd', 
tvdyd' 3, tudyd' 8, dansdnd 7, dakshind 1 2, dacasyd', dindtd, duvoyd', devatd i.22.5; 
55.3; 100.15: vi.4.7. devayajyd' 2, doshd' 13, dvitd' (? Roth, ' so fort, fortan, von 
nun an') 29, dhanyd and dhishand, v.41.8, dhd'rd, dhishd' (see Gr.) 2. pavd', 
pdkid 3, pwrushatd 3, purushatvatd 2, bandhutd, oarhdnd 13, brahmanyd' (see 
above), bhand&nd 2, rnanhdnd 11, madhyd' 4, manand', l.mand', l.mand', manUhd' 
13 (in iii.57.1, Aufrecht' 2 has -d'm; Mflller, -d'; the scholiast seems to have read 
-d'm), mamatd, mr'shd, mehand 4, yavid' 2, rathayd', (vakshand, v.52.15, Say., 
vahanena nimittena; better N. ?) vacasyd' 3, vapushyd', va,rasyd', yarivasyd', 
vasutd, vas&yd' 2, vipanyd', viprd (mati' viii.25.24), virayd' 2, redatd, vedid', 
vedhasyd', cdsd (Say., stutyd, v.41.18), praddhdrnanasyd', craiiasyd' 7, sanatd 2, 
samand.' 13, savyd', sasvdrtd, sukrtyd' 2, sukraluyd' sukshetriyd', sugdtuyd', 
swjdtatd, sumnayd', supravasyd', susanitd, svadhd' (see below) 2, suapasyd', himid', 
hiranyayd'. To these we may add namasyd', ii.33.8 (so A. Weber); dhend, 
x.104.10; fvetid', x.75.6. 

The Atharvan has some of these instrumentals in Rik-passages : 
three in iv. 33.2 ; one in xviii.1.52, and one in xix.12.1; and some 
in passages peculiar to itself: the stereotyped g-uhd, 13 times (and 
twice in Rik-passages) ; dakshind', ix.7.20 and xii.2.34 (but 
ddkshind in the Rik-passages, xviii. 1.42 and xix.13.9!); devatd, 
iv.1.5; dosha', vi.1.1 and xvi.7.9 ; sumnayd', vii.55.1 ; and 
vittakdmyd' (not a Rik-word), xii.3.52. So kimkdmyd' , Qat. Br. 
i.2.5" and yatkdmyd', iii.9.3*. 

2. Examples of the common form are : jihvdyd, 24 ; dhd'rayd, 53 ; mdydyd, 20 ; 
samayd, 9; hrdayydyd, x.151.4. 

There are several forms bearing more or less resemblance to an I.s.f. of the 
pronoun a: ayd, x.116.9, p. dyds, is N.p. of stem dya, 'wanderers;' possibly ayd 
(so Aufrecht* and Codex Chambers 60), vi.66.4a, p. ayd, is from the same stem, 
for dyds — 'as long as the ones now wanderers quit not their birth-place;' in 
vi.66.5a, text ayd's, the cadence requires dyd's, i. e., asyd's (Gr.), Ab.s.f. of a; 
ayd' occurs as substantive pronoun in ii.6.2 (sc. samidhd), as adjective pronoun 
with substantive expressed, 18 times (6 with dhiyd'), and as adverb iii.12.2: 
vi.17.15: ix.53.2; 106.14, and, I must add, i.87.4J (cf. Prat, ii.29: Miiller, 
Translation, p. 148). For andyd, ix.65.12, 27, Gr. reads ayd'; but anena, i.93.10: 
iii.17.2 : vhi.44.2, warrants us in holding fast to the text. 

Dative Singular Feminine. 

This case terminates in -dyai. It occurs only 20 times (from 15 
stems). The full list is: agdtdyai, avi'ratdyai, ghoshdyai, 



D.s.f.] Noun- Inflection in the Veda. 359 

card'yai (Delbrtlck, in Kuhn's Zeitschrift, xviii.85), jard'yai, 
tvdyatdyai, duckkndyai, p-Mdkratdyai, mand'yai, ii.33.5 (in 
iv.33.2, Roth takes it as G.s.f., q. v.), vippdldyai i.116.15 only, 
givd'yai, pvetand'yai, stinr'tdyai 2, sdrid'yai 3, sud'yai 2. For 
prajayai see below. 

In i.54.11, we have suapatyai (stem suapatyd' ; scholion to 
T. Br. ii.6.9 1 , pobhandpatyatvdya), and in i.113.6, mahiyai (stem 
mahiyd' ; see BR. vii. 1789). They are not for suapatyd'yai and 
mahiyd'yai; but rather relics of the simpler formation, the general 
ending (here e) being added directly to the stem, as with the 
instrumentals, p. 357, and the infinitives khyai, dai, mai, yai, sat 
(Delbrtlck, Verbum, p. 221). 

In vii. 1.19a, we read md' no agne avi'rate pdrd ddh; perhaps 
this form was preferred to md'vt'ratdyai no agne pdrd ddh, which 
would be metrically correct, because in the latter the sandhi 
would obscure the a- privative of avi'ratdyai. In iii.16.5, md' no 
agne dmataye : md'vt'ratdyai riradhah, this obscurity is remedied 
by the parallel dmataye preceding it. The genuine Vedic hymns 
are as averse to ambiguity as some writers of the later literature 
are fond of it. We have here a simple transition to the consonant 
declension ; cf. devdtdtaye and devdtdt-e, sarvdtdtaye and Zend 
haurvatdt-em. 

In i.118.8, prdti jdnghdm vippdldyd adhattam, p. -as, 'Ye put 
on again Viepala's leg (that was cut off),' sense and form (Prat. 
ii.9) admit of a dative ; ' Ye put on a (new iron) leg for V.,' as in 
116.15. 

Ablative Singular Feminine. 

The ablative and genitive s.f. are coincident in form, and 
terminate in -dyds. The separation here is based on the exegesis. 
The clause, x.87.16, yd aghnyd'yd bhdrati kshirdm agne, is a good 
example for showing how indistinct the logical border-line 
between these two cases is. The ablatives are as follows (8) : 
kand'yds, jihvd'yds, ddkshindyds, durevdyds, durhdndyds, 
daividyds, niddyds, sud'yds; and in AV. iv.40.2, ddkshindyds; 
5, dhruvd'yds; 6, vyadhvd'yds; 7, tirdhvd'yds. In ayd's, RV. 
vi.66.5a, Gr. sees an Ab. See p. 358, and cf. ddkshindyd[s], G.s.f. 

Genitive Singular Feminine. 
^ The genitives number 54 (from 27 stems) and are as follows : 
dghnidyds 3, aghnyd'yds, iddyds 5, ukhd'yds, usriydyds 5, iX'rmid- 
yds 2, kand'yds 3, kd'shthdyds, jihvd'yds 2, ddkshindyds 4, 
darpatd'yds, dU'rvdyds, dhishdndyds 2, dht'shamdndyds, pdjrdyds 
(with voc), pdritakmidyds, md'hindyds, rasd'y'ds 2, vayd'yds, 
vi?pdldyds (i.118.8), papvattamd'yds, piphdyds, pucd'yds, sabardu- 
ghdyds 2, sehdnd'yds, s&rdyds, stirid'yds and sUryd'yds 8. 

In i. 123.5, jdyema tarn ddkshinayd (p. -d) rdthena, one is 
tempted read ddkshindyds in view of la, and to interpret, with 
Gr., as Gen., the d being due to the metre; but cf. x.26.6a 
and 107.1a'. BR. take mand'yai (s.v.) as G. in iv.33.2, 



360 C. B. Lanman, [a-stems. 

dhi'rdsah pushtim avahan mand'yai. I find no form in the Veda 
to support it, although the case is so extremely common in the 
Brahmanas : Say., mandyd rtena; Gr. ' zum Lohne.' CI'., on the 
other hand, vifpdldyd(s) as D.s.f. before a-. 

Locative Singular Feminine. 

This case terminates in -dydm. It occurs 30 times (from 18 
stems). Scherer, Ges. d. Deutschen Spr. 283, and Bollensen, Z. D. 
M.G. xxii.618, speak of locative forms in -d from d-stems. I do 
not think that any such can be proved. Some of the supposed 
ones are instrumentals (Kuhn, Zeits. xviii.365, and note to 367). 
Benfey (SV. ii.830 ; Ueb. p. 278) sees a L. in rasd' dadhita 
vrshabhdm, RV. viii.61.13; but dadhita is 3 s. opt., and rasd' its 
subject : Gr., 'let the flood receive the steer.' 

The forms follow : dpayd'ydm radical ?, dmd'ydm, uttdnd'ydm 
2 and i.164.14, usriydydrn 3, -gatdydm 2, grivd'ydm, jUrnd'ydm, 
ndvdydm, pdritakrnidydm 5 and with djdl m., i.116.15, bhadrd'ydm 
2, yamtindydm, varatrd'ydm, pinfdpdydm, pirindydm, sti/rdydm, 
sushdmdydm, svd'ydm 2, hariytipi' ydydm. 

Vocative Singular Feminine. 

The V.s.f. ends in -e. It occurs 95 times (from 51 stems). If e 
represents an d, the change is purely phonetic. The Rik shows 
no trace of a form in d. The " still weaker form in d " (Kuhn, 
Zeits. xviii.369 and Schleicher, Comp.* p. 574, § 263) is supported 
by amba only, ii.41.16 and x.86.7 bis; and it is doubtful whether 
in the Veda amba means 'mother' at all. In x.97.2 it is surely a 
particle. Like akka and alia, given in the scholion to Pan. 
vii.3.107, amba is probably a natural cry fitted into a grammatical 
category ; so BR. By the side of this form, as if it were from a 
real stem amba, arose the word ambe,VS. xxiii.18 — scholion: 
patnyah parasparam vadanti 'he ambe.^ 

The forms follow: aghnie, apue, amartie, ambitame, apve, 
apvasdnrte 10, drjikiye, ide, ugraputre, uttare, uttdnaparne, 
uruvraje, tXrraie 2, rshve, kadhapriye, kdne, gauge, ghore, citre, 
citrdmaghe, jdye, devajtite, devitame, dhishane, naditame, pathie, 
pastie, putrakdme, purupriye, prthujdghane, prthushtuke, 
brahmasampite, marudvrdhe, yarnune, rdke, vikate, vipvavdre 3, 
paravye, <$tike, pubhre 3, saddnue, sarame 4, site, sujdte 13, 
suputre, subhage 12, suldbhike, susnushe, sflnrte 4, sUrie and s-drye, 
hlddike (Gr., Ueb.; in Wb. as L.). 

Nominative, Accusative, and Vocative Dual Feminine. 

These cases are coincident in form, and end in -e. They pccur, 
taken together,-321 times (from 136 stems). The termination is 
of obscure origin. It is pragrhya (Prat, i.28), and the pada 
distinguishes these dual forms from the vocatives singular by an 
iti. Of. ubhe iiti) ydt te mahind' puibhre dndhasi, vii.96.2, with 
$ubhre (iti), iii.33.2; i.30.21c with iii.33.16; and x.l4.16<? (yame, 
L.s.m.) with 13.2a. 



N.A.V.d.f.] Noun- Inflection in the Veda. 361 

In ii.3.6, ushd! sdndktd vayieva (p. -d-iva) ranvlt'e, and iv.32.23, 
kaninakeva (p. -d'-iva) vidradh'e, the feminine adjectives require 
us to read vayie va and -Ic'e va. So we may read, with Bollensen, 
ii.39.6, nd'se va (p. -d-iva), in view of these two instances, although 
nd'sd might be taken as N.d.f. from stem nds, " strong " nd's. 

In i.29.3, sastd'm dbudhyamdne, Gr. reads -mdnai. This is 
hardly to be allowed. The pdda is catalectic (see p. 338, top). 

Examples are : ubhe, 66 ; devdputre, 7 ; yam'e; virdpe, 7 ; pipre, 
5 ; sumeke, 8. 

Instrumental, Dative, and Ablative Dual Feminine. 

In x.105.5, piprdbhydm seems to be an instrumental. Nd'sih- 
dbhydm, x.163.1, is an Ab. In x.88.15, td'bhydm is I. with 
distinct locative meaning. Cf. p. 343. 

Genitive and Locative Dual Feminine. 

The genitives are: ayds i.185.1; ydyos ii.32.1 and iii.54.2; 
yamdyos x.8.4. For janghdyos, AV. xix.60.2, see below. 

The locatives are: ayds (sdcdydh, p. dydh) iii.54.2; uttdndyos 
i.164.33; ydyos iii.44.3. For svadhdyos, ix.*86. 10, see below. 

Nominative and Vocative Plural Feminine. 

The N.p.f. has two forms. The almost universal termination, 
-as, occurs 586 times (from 262 stems), thus coinciding with that 
of the A.p.f. ; and, in general, these two cases are not formally 
separated in the Veda. But the ending -dsas is also found as 
N.p.f. some twenty times, more or less. 

The comparative grammars make much of the feminine forms 
in -dsas, and set up entirely different organic forms for the N.p.f., 
e. g., *dg,yd-sas, and Ap.f., e. g., *dpvd-ms. But the evidence 
upon which this assumption is based, so far as it comes from the 
Veda, is slight. I would quite as lief consider the feminine forms 
in -dsas an irregular extension of the use of the masculine form to 
a field where it did not belong. Some of the verses in which 
these forms occur seem to be of late date, although this can hardly 
be affirmed of them in general; such are iv.58.7, 8 : vi.63.9: 
x.95.9: and perhaps v.46.7 : viii.44.25; 82.23. 

The evidence of the kindred tongues, on the other hand, seems, 
as far as I can judge, to speak for the originality of -dsas in the 
feminine. I can find no N.p.f. in -donhd in the Avesta ; but see 
below, A.p.f., p. 363, and Bopp, Vgl. Gr.' i.460, §229. The 
Prakrit N.p.f. ending in -do represents, I believe, an original -dsas, 
although, to be sure, it is seen also in the A.p.f. (Prdkrta-prakdpa, 
v. 19, 20); thus, pamaddo, Cak. 4, stands for pram'addso. The 
Pali N.p.f., kanndyo, stands probably for kanfido, with euphonic 
y (see Kuhn, Pdli- Gr. p. 64), and this' for kanydso. The materials 
for verifying these statements, however, are not at my command. 

1. Examples of the common form are : imd's, 32 ; etd's, 7 ; td's, 
47 ; dhd'rds, 16; bhadrd's, 19; mantshd's, 10; yd's, 71 ; vayd's, 
8. Besides the 586 forms mentioned above, vipvds, N and A, 



362 C. B. Lanman, [^-sterns. 

together, occurs 104 times. Vdyd's, x.92.3, has been counted 
under vayd', ' twig.' 

Among the 586 forms there may be some vocatives. I have 
noted only two separately, vi$vaminvds and suhavds. 

The pada has -ds in vi.29.2, where we must read ndrid: see 
p. 348. 

In vii.6 1.5, on the other hand, we have to read drrvdrd{s) vipvd(s), 
p. -d. Say. seems to have had MSS. that read imd' (imdni). 

Elision and crasis is seen in ix.96.24, d' te rdeah pdvamdnasya 
soma ydsheva yanti sudtighdh sudhdrd'h, i. e., y6shds-iva, p. -d-iva. 

The resolutions, usrdas viii.64.8, br' haspdtiprasAtaas x.97.15, 19, 
samviddndas 14, p&nyagandhaas vii.55.8, are all at the end of 
catalectic pddas of seven syllables, and unnecessary. See p. 338, 
top. In x.26.1, pronounce spdrhd! idnti niytitah, not spdrhdas 
(Gr.). For vayd as, vii.40.5, see the end of this page. 

2. The following is an enumeration of the N.p.f. forms in -dsas, 
including at the end some doubtful ones. To show that the native 
authorities recognized them as feminines, I have added a few 
excerpts from the scholia. 

A'tandrdsas, durmitr&'sas, pd'rthivdsas, pdvakd'sas, bhejdnd'sas, vanvdnd'sas, 
vafd'sas ' cows,' vdfrd'sas, viddnd'sas, vrdhd'sas, c4ghand'sas, smdyamdndsas, 
dfushdnd'sas, pasprdhdnd'sas, hdvamdndsas, dfvdsas ?, janydsas ?; A V. sdmmitdsas, 
amr'tdsas, vafd'sas, martdsast. 

i.95.2 (TBr. ii.8.7 4 ), dtandrdso yuvatayas. vii.28.4, durmitrd'so hi kshitdyah 
pavante, ' denn die bosen Leute miissen biissen,' Roth. v. 46. 1, yd'h pd'rthivdsas 
. . . td' no devih fdrma yachata; Say., prthivtsambandhinyah. i.142.6, pdvakd'so 
dvd'ro devt'h; Say., fodhayitryah. x.153.1, Inkhdyantir indram ii/pdsate ihejdn- 
d'sah, sc. indramdtarah, Anukramani. The SV., i.175, has vanvdnd'sah. 
vi.63.9, fdndo \a\ddd dhiraninah smdddishtin, ddfa vafd'so dbhishd'ca rshvd'n, 'Q. 
gave horses with golden trappings and well-trained : ten cows (mares ?) follow the 
mighty (steeds).' So Both; but Say., ' vagd'so' 'nugundn ' dbhishd'cah'' gatrdm&m 
abhibhdvukdn ! viii.44.25, giro vdfrd'sa irate; Say., vdfanafildk. i.169.2, 
dyujran . . vicvdkrshtir viddiid'so nishshidhah. viii.82.23, hotrd asrkshata indram 
vrdhd'sah; SV. i.151, vrdhantah. iv.58.7 (VS. xvii.95), cughand'sah patayomti 
ghrtasya dhd'rdh ; Say., dcugantryah. iv. 58.8, yoshdlt- smdyamdndsah : Sehol. to 
VS. xvii.96, ishaddhasantyah. iv.24.4, kshitayo . . dpushdnd'sah. vii.18.3, 

imd' » tvd pasprdhdnd'so atra mandrd' giro devayantbr upa sthtji, 'And unto 
thee pressed onward then our emulous, gladsome, godly songs.' v.32.11, tarn 
me jagrbhra df&so dosha' vdstor h&vamdndsa indram, ' On him my hopes laid hold, 
which unto Indra morn and eve do cry.' 

The following two are possibly masculine : x.95.9, acvdso na kriddyah: Roth, 
Fsrl. zum Nirukta, p. 154, note, 'wie lustige Stuten.' ix.49.2, gd'vo janydsah; so 
SV.ii.786. 

The Atharvan has several : two in xii.3.27, uteva prabhvi'r utd sdmmitdsa uta 
cukrd'h cucayac cdmr'tdsah, sc. d'pah; and one in iv.24.4, ydsya vacd'sa rshabhd'sa 
ukshanah, 'to whom belong cows, steers, and oxen,' as in Rik vi.16.47. In 
xviii.3.23, martdsac cid v/r\u~\vacir akrpran, unless we read mdrtydsac cid, the 
preference is doubtless to be given to the Rik reading, martdndm cid, iv.2.18, 
since marta is never an adjective. Pd'rthivdsas, vii.49.1, TBr. iii.5.12 1 , is merely 
quoted from the Rik, v.46.7. 

In the Rik, x.91.14, avasrshtd 1 sas agrees with nouns m. and f. 
In vii.40.5, vayd'h has to count as an amphibrach, vayd' sas; but 
Gr., Ueb., proposes vaydma. Cf. as-stems, A.s.m. 



A.p.f.] Noun- Inflection in the Veda. 863 

Accusative Plural Feminine. 

This case ends in -ds. It occurs 393 times (from 163 stems). 
Vipvds, moreover, N. and A. together, occurs 104 times. Vdmd- 
jdtds, x.140.3 (Gr., N.p.m.), is included here; see i-stems, A.p.f. 

1. Examples are: dmivds, 7; imd's,25; usrd's,12; usriyds, 
10 ; ends, never nominative, 9 ; if an accented accusative-form is 
required, etd's is used, 5; td's, 17; dhd'rds, 11; pr' tanas, 13; 
mdyd's, 22 ; yd's, 15 ; silnr'tds, 6. 

The pada has -d in a number of instances where the reading -ds 
is more or less certain : drnd{s) sisarshi, iii.32.5 (see p. 348) ; 
dhd'rd(s) ?, ix.7.2 : patkid(s), vii.79.1 ; pdrvi'r mamshd'(s), i.70.1 ; 
vipvdh should be added to the jo«<fo-excerpts for x. 153.5c, 
Aufrecht, Migveda*. For pratishthd' (s) hr'did(s), x.73.6, see 
stems in d radical, I.s.f. For cdsd'(s), vii.48.3, see stems in s 
radical, I.s.f. For ushd's, see as-stems, A.p.f. For gnd's and 
maryd' dds, see stems in d radical, below. 

As regards resolution — it is not necessary to read manishdas, 
x.26.1a, nor usrdds, viii.46.26a, since both pddas may be catalectic. 
On the other hand, three seem hardly avoidable : kdas, vii.60.95/ 
kd'shthaas, i.63.5 ; nv&rdds, iv.26.7. 

2. If -dsas is to be allowed at all as ending of the A.p.f., it is 
merely as a solecism. In x.30.14, ni barhishi dhattana somidso, 
apd'rh ndptrd sarhviddnd' sa endh, ' Set upon the barhis, ye soma- 
oflferers, the (waters) in concord with Apam Napat,' we cannot 
make endh subject of a separate clause, since etd's is the N.p.f. to 
which ends serves as A.p.f., as in x.19.2, 3; nor can we, taking 
sarhviddnd' so as N.p.m., make endh the object of n'% dhattana, 
since it is improbable that an unaccented word would stand alone 
and separated from its verb by a long parenthesis. It only 
remains to take sarhviddnd' sas as A.p.f. with endh, sc. apds, and 
this is favored by verse 13, sarhviddnd! d'pah, (So Say. takes 
vapd'sas, vi.63.9, as accusative.) But this hymn is so good that I 
am half tempted to get rid of this unconscionable solecism by 
reading etd's for ends, or else sarhviddnd' apainds, i. e., -nd's apds 



If, however, we let it pass, it is paralleled by AV. xiii.2.33, 
where the sun is spoken of as ' furnishing ready paths ' by its 
light — ararhgamd' sah pravdto rdrdnah. Here there seems to be 
no question about form or sense. 

Ghendo hufedhris huraodhdonhd dyepe, Vispered ii.7, 1 trans- 
late thus : ' gndh supitfh sur&dhdsa dydce,' and take huraodh- 
donhd as A.p.f. from the stem huraodha, rather than set up 
another stem for this one form, as does Justi. 

Bopp's two accusatives p.m. in -donhd, Vd. xviii.65, rest upon 
misinterpretation; they are N.p.m. ; so mashydonhd, Yc. xxx.ll, 
is V.p.m. See Vgl. Gr. 3 § 229. 

Instrumental Plural Feminine. 

The general ending -bhis is affixed directly to the stem : as, 
d'bhis. The form occurs 295 times (from 89 stems). Examples 



364 0. R. Lanman, [d-stems. 

are : iddbhis, 6 ; citrd'bhis, 8 ; td'bhis, 50 ; dansdndbhis, 5 ; dhd'- 
rdbhis, 7 ; mdyd'bhis, 13; yd'bhis, 50; vipvdbhis, 23; vedid'bhis, 
5 ; pivd'bhis, 5 ; silnr'tdbhis, 5 ; hdtrdbhis, 7. 

In iii.62. 17, drd' ghishthdbhis is used adverbially ; cf. apart' bhy as. 
Monosyllabic stems suffer no change of accent : thus, td'bhis, 
yd'bhis, svd'bhis; but from a we have dbhis (4), dMis (10), 
d'bhis (3). 

Dative and Ablative Plural Feminine. 

The general ending, -bhyas, is affixed directly to the stem : as, 
td'bhyas. It sometimes counts as two syllables. The two cases 
are of course coincident in form, but I have grouped them here 
according to the exegesis. 

Datives: aghnid'bhyas, v.83.8 ; dbhyds, ii.32.2 ; usriydbhyas, 
vi.IY.6; td'bhyas, x.169.2 ; yd'bhyas, vii.47.4; vrtd'bhyas, 
x.30.7: — dbhias, vi.30.3 ; duchdndbhias, ii.32.2. 

Ablatives: dbhyas, x. 17.10; ushnihdbhyas, M'hasdbhyas, and 
grtvd'bhyas, x.163.2 ; td'bhyas, x.19.8 ; sdrvdbhyas, ii.41.12: — 
ddhardbhias and tittardbhias, x.145.3; d' (dbhias, ii.41.12; 
ydddbhias, x.163.3; dansdndbhias, iii.3.11 ; dhishdndbhias, iv.36.8: 
ix.59.2; yd'bhias, vi.47.3 ; vakshdndbhias, i.134.4 bis; viii.1.17; 
pydvidbhias, vi.15.17. 

Genitive Plural Feminine. 

This case is not distinguished in form from the masculine. It 
occurs 39 times (from 22 stems). The metre requires a short 
thematic a in one instance, i.3.11, sdnrtdndm (cf. p. 352). The 
G.p. of kanid, *kamdndm, always appears in a contracted form, 
kani'ndm (five times). 

1. Enumeration: dksharandm, Aghnidndm, anivefand'ndm, tibhaydndm, apma- 
vrajdndm, urv&rdijAm, usrd'ndm, usriydndm 1, kd'shthdndm, jihmd'ndm 2, divy- 
d'ndm, dughdndm, devasend'ndm, dhishdndndm, n&vdndm, ndvidndm 2, pastidndm, 
3, mantshd'ndm, rd'midndm 2, samdndm, s'Umr'tdndm (see above) 6, stiydndm 2. 

2. Gr. proposes resolution in eight instances: in viii.59.12, 
dhdnd'naam, at the beginning of a pdda; inviii.59.1, pr'tandnaam, 
thus converting the cadence from trishtubh to jagati: — in the 
other six it is of questionable necessity, at the end of pddas of 
seven or eight syllables: dghnidnaam, viii.58.2 ; dntamdnaam, 
i.4.3; iddnaam, ix.108.13; pastidnaam, ix.65.23; vedid'naam, 
x.22.14 ; sUnr'tdnaam, viii.32.15. 

3. There is no certain case of the simple ending -dm, without n. 
We find ndirikam, Yt. xxiii.l, and gadhwam, Vd. xv.16, 17. If 
Gr. (Ueb., note to ix.12.7) is right in following the Sama variant, 
dhend'm dntar sabardHghdm, ii.552, then the two words must be 
taken as accusative, and not, with Benfey, Glossar, s.v. dntar, as 
genitive. 

Locative Plural Feminine. 

The general ending -su is affixed directly to the stem : thus, 
td'su. Forms of this kind occur 182 times (from 61 stems). 
The final u is never combined with a following initial vowel. It 



L.p.f.] Noun- Inflection in the Veda. 365 

occurs at the end of a pdda before a vowel in iii.30.14; 55.9: 
vi.65.1 : x.27.16. It occurs in the interior of a pdda: before a-, 
14 times, as ii.40.2 ; before d-, iv. 2.12 ; i-, i.127.6 bis: x.147.2; i-, 
iv.51.7 ; u-, x.50.3, he apsii, svd'su urvdrdsu patinsie; H-, vii.49.45/ 
r-,i.l60.4; before the word «', i.25.10 ; 141.5: vi.48.6 : viii.39.7 : 
x.49.10. In all these 31 instances the ending is written as if 
combined, but is to be pronounced with hiatus. The result of this 
examination corresponds entirely with that above, p. 354. 

Enumeration : aghd'su, anyd'su, amr'ldsu, avardsu, dmd'su 6, d'artandsu, iddsu 
2, dsu 16, dsu 9, upardsu 2, urvdrdsu 5, usriydsu 2, u'rmidsu, kd'su 4, katamd'su, 
kanidsu 3, kd'shtdsu 2, krshnd'su 3, jagmdnd'su, cittagarbhdsu, citrd'su, jdtd'su, 
td'su 2, tugridsu, duridsu 4, dhishnidsu, dhruvd'su 2, ndvdsu, pathidsu, pddydsu, 
pastidsu 3, pd'ncajanydsu, purupefdsu, p&'rvdsu 2, pr'tandsu 35, pradhanidsu, 
priyd'su 2, madidsu, madhyamd'su, marmshidsu, mdrtidsu, mahin&su, yd'su 12, 
yajiiiydsu 3, -yatdsu, yoshandsu, rd'midsu, ropand'kdsu, vakshandsu 6, vipvdsu 16, 
vrdhasdnd'su, vrddhd'su, (ayd'su, pushkdsu, fydvd'su, prutd'su, sdnaydsu, saptdfivdsu, 
sird'su, suvrjandsu, svd'su, sud'su, havidsu. 



STEMS IN / AND 1. 

The Vedic noun-inflections differ more from those of the later 
language, and are historically more clearly separated from them, 
upon the field of the i and ^-declensions than any where else. The 
stems ending in i and i exhibit also in the Veda itself a greater 
variety of declensional forms than any others, although the u and 
■^-declensions have had a closely parallel development. The 
reason of this variety is to be sought in the physiological 
character of the stem-final ; this lies on the border-land between 
vowels and consonants, and is the one or the other according to 
circumstances. (See Professor Whitney's lucid exposition of this 
subject in the Journal of the Am. Or. Soc. viii.361,362.) Thus 
the instrumental ending -d can be affixed directly to a thematic i 
with hiatus, giving the termination -id; or this i may take on a 
consonantal character, so as to result in -yd; a very common 
phonetic change reduces either of these groups to 4; and this may 
even be shortened to -i; or, finally, a n may be "inserted" 
between stem and ending, so as to produce a fifth form -ind. In 
no case, however, except -i of the dual, does thematic i (I) coalesce 
with an initial vowel of the ending (as would a thematic «) ; but it 
sometimes absorbs it, About one half of the case-affixes begin 
with vowels ; and in these cases, according to the rules of Sanskrit 
writing, the preceding thematic i becomes y; but the evidence of 
the metre shows that in the Vedic language this mutation was by 
no means accomplished. The simplest and most organic form of 
the I.s.f., for instance, would seem d priori to be -i-d, and this it is 
also in fact. As in the later language -yd is the exclusive form, 
so in the Veda -id predominates. The same is true, for the most 
part, of the other case-forms with vocalic endings. But we see a 
historical movement in progress in the Vedic texts ; and the forms 
in which the thematic vowel has become y belong in general to 
vol. x. 50 



366 C. R. Lanman, \i and 

later hymns. Thus, in the G.L.d.m. of Si-stems we have 30 forms 
in -ios and none in -yos, and in the Ls.f. we have 37 forms in -id 
to 13 in -yd; but in the secondary forms of the D.s.f. we have 3 
in -iai to 6 in -yai. 

A careful study of the mass of forms seems to point to three 
sets or series of endings : 



ries A. 


Series B. 




Series C. 


i s 


id 


i 




V s 


i m 


id am 


i am 


im 


V am 


i d' 


id d' 


i &' 




i' d 


e e 


id e 


i ai 




V e 


e s 


id as 


i d's 




V as 


e i 


id d'm 


I d'm 






e 


id 


i 




i 


i i 


id i 


i i 


i 


V d 


i bhydm 


id bhydm 


% bhydm 




V bhydm 


i 6s 


id 6s 


i 6s 




V os 


e as 


id as 


% as 


is 


i' as 


i ns 


id as 


i as 


is 


V as 


i bhis 


id bhis 


i bhis 




V bhis 


i bhyas 


id bhyas 


i bhyas 




%' bhyas 


in d'm 


idn d'm 


in d'm 




i'n dm 


i shu 


id su 


i shu 




V shu 



It will be seen that series A and C differ most from each other. 
Series A is a true vowel-declension ; series C, rather a consonant- 
declension. This is as it should be. A is used with stems ending 
in a short vowel, which is more readily susceptible to guna than 
a long one ; C is used invariably with stems ending in long i, and 
that i is always accented and never gunated. For the very 
reason that it is accented, the individuality of the vowel is more 
perfectly preserved. It never passes into y in the Rigveda, while 
the thematic i of series A very often does so. Series C never 
shifts the accent to the ending, while A does this regularly, if the 
stem is oxytone (exceptions will be noted). 

Series B, genetically considered, lies between A and C. Com- 
parative grammar teaches that the % of feminines with which B is 
chiefly used was developed out of a fuller derivative ending id. 
Cf. para, pdrya, ydryd, $d'ri; *dd-taryd, ddtrt, *dorepya, 
doreipa; avit(a)ri'; and see Schleicher, Comp. 4 p. 382 §217, p. 
390 § 218, etc. Even with the feminine <?-stems we saw some 
relics of simpler modes of declension : as, in the instrumental (pp. 
357, 358) ; in mapatyai, maMyai (359) ; but here the whole series 
may be developed phonetically from the combination of id with 
the endings of C, except N.A.d. Before -d, -e, -as, and -dm of 
the singular, the thematic d united with the vowel of the ending ; 
elsewhere with the preceding i to i. In the A.s., N. and A.d. 
and p., the resultants suffered further contraction : 4am became 
-im; -ii, -i; and -ias, -is. The resulting series of terminations was 
established as follows : -i, -im, -id, -iai, -ids, -idm; -%, -ibhydm, -ios; 
-is, -is, -ibhis, -ibhyas, -indm, -ishu; and of these terminations, the 
case-suffixes were apparently — , m, d, ai, -ds, dm; — , bhydm, os; 
s, 8, bhis, bhyas, ndm, su. 



2-steni8.] Noun- Inflection in the Veda. 367 

It is evident that in the case of oxytones the I.s., G.L.d., and 
G.p. of series B and C would be coincident, were it not for the 
fact that B shifts the accent regularly to the ending, while C 
never does so. 

A mutual relation exists between the place of the accent in the 
weak cases singular of oxytone 2-stems and the strengthening of 
the stem, either by vowel-change or by " addition " of a nasal. 
It is quite analogous to a like one in verb-inflection. Thus, in the 
I.s., if the stem remains weak, the accent is shifted to the ending: 
as, pavy-d! (cf. i-rnds) ; so indrdgnids; but if the i is gunated, it 
retains the accent : as, agndy-e (cf. e-mi) ; in like manner, tir/iii-d', 
tirmin-d; ari-ds, agn'e-s. 

Series A was applied to the stems in short i, including many 
masculines and feminines. and very few neuters (only 322 n. forms 
occur). It was applied originally to these stems without distinc- 
tion of gender — except, of course, that the neuters had in the N. 
and A. no ending in the singular, 4 in the dual, and -i in the plural. 
At any rate, the distinctions of the grammarians have no existence 
in the Rigveda. (Cf. Bohtlingk, Die Declination im Sanskrit, 
Mbnoires de VAcademie Imperiale des Sciences de St.-Peters- 
bourg, vi me Serie, T. vii. p. 164.) Here belong further a few 
stems which appear to end in a root ; they are the compounds of 
-dhi (BR. iii.959) and others: as, d'dhi, sddhi, pradhi; pratishthi; 
yayi from yd; tuvigri. Lindner, p. 56, supposes that the root 
has dropped its final and taken the suffix i; but has not the final 
& been weakened to i ? Declensionally, they are treated entirely 
as stems ending in suffixal i (cf. the accent, nidhind'm) and are 
placed here accordingly. 

Series B was applied to a large class of stems, almost all 
feminine, ending originally in id and contracted to i. These 
stems form feminines to words requiring inflection in more than 
one gender, and correspond for the most part to masculines of the 
following groups : many words in a (e. g., devt) ; adjectives in 
u (dnvi, dpipvi, vdsvi, and urvt', trshvi', ptirui', prthvi , prabhvl', 
bahvi', yahvi' ?, raghvt' , vibhvi', sddhvi', and svddvf) ; participles 
in -ant (mddanti), -dnt (adati 1 ), or -at (piprati) ; and -vans 
{jagnvdsht) ; comparatives in -iyans (ndviyasi) ; words in -tar 
(avit[a\ri'), in -vant {dmavatt) and -mant (dhenutndti), -van 
(rtd'vari), -an (compounds of H'dh[a\n, pirsh[d~\n), and -in 
(arFmi) ; many compounds of ac (arvd'ci) ; some of han 
(dpatighni), drp (sudr'pi), and pdd (apddi). Some of these 
have double feminines (apdd; pdpd' and pdpt' ; etc.). Besides 
the above, there are about 120 Rik- words belonging to this 
declension, which are of very heterogeneous character, as 
regards their formation. They are the sterns in i given by 
Grassmann, columns 1722 and 1723, excepting those that I give 
below as belonging to C. To these must be added finally seven 
stems, designations of males, but declined like the feminines of 
series B : Tirapei', JSTdmt, Dr'thi, Md'tali, and S6bhari (all proper 
names) ; rd'shtrt, ' ruler,' and sir?, ' weaver,' 



368 C. R. Lanman, [i and 

Feminines formed with change of accent are declined according 
to C, if the resulting stem is oxytone. I therefore suspect that 
the accent of asiknid', x.75.5 (elsewhere dsikni, dsita), is erroneous. 
It may be that the peculiar accent here differentiates the word as 
a proper name, for it denotes a river of the Pafijab (cf. (Jdyamdnd, 
vi.27.8; Citra, viii.21.18; Didyut, AV. ii.2.4, an Apsaras; so 
hastini ' having a hand ;' but hastini' ' female elephant,' AV. 
vi.70.2) : even then we ought to write asiknid; so for sddharanyd' 
(sd'dharana), i.167.4, read -yd". (Japvati', viii.49.17, and gomatt, 
twice, are also incorrect. 

Barytone feminines, on the other hand, corresponding to 
oxytone masculines, are declined according to B; they are: 
drushi, d'yasi, gd'ndharvi, tdvishi, pdrushni, pdliknl, mdhishi, 
vdrdtrt, pamgdyi, pyd'vi, py'eni; and rohini, hdrini ? 

Oxytones of series A and B will be enumerated in full, in 
cases where the accent is thrown forward to the ending. 

Series C belongs to a comparatively small number of words : 
(1) properly to stems whose final element is a root ending in i! 
(grdmant, m., f.) and (2) to monosyllabic -substantives (f.) and 
their adjective compounds (m., f.) in i'. In like manner are 
declined (3) about fifty oxytone feminine Rik-words in *', of which 
all but half a dozen are substantives. Of these («) 22 correspond 
to barytones (mostly masculines in a), or are derived from them 
with shift of accent, and all but five designate female animate 
beings; atharvi' (dtharvan) 'priestess;' ardyi' (drdya) 'demon;' 
ashtakarni'. (ashtdkarna ?) ' cow with an 8 on her ear ;' eni' (eta) 
' doe ;' kalydni' (kalyd'na) ' fair woman ;' kildsf (kild'sa) ' spotted 
deer;' napti' (ndptar) 'daughter;' purushi' (ptirusha) 'woman;' 
mand&ki (mandti'ka) ' female frog ;' rnayUri! (mayU'ra) ' pea-hen ;' 
ydiudhdnt (ydiudhd'nd) ' sorceress ;' rathf (from rdtha) ' female 
charioteer;' laldmU (lald'md) 'speckled mare;' vipvarilpi' 
(vipvdrilpa) ' brindled cow ;' vrki' (vr'ka) ' she-wolf;' sumangali' 
(sumangdld) ' lucky woman ;' d-durmangali ' not unlucky ;' [add 
from TS. : pabali ($abdld) ' die Wunderkuh ;' mahishi' (Rik 
mdhishi, mahishd) 'buffalo cow;'] further, apart (dpara), pi., 
' future days ;' tapani' (tdpana) 'heat ;' $akati! (pdkata) ' cart ;' 
sakthi' (sdkthi) ' thigh ;' srni' (sr'ni) ' sickle.' ' (b) Twelve others 
denote female animate beings, but do not correspond to barytone 
masculines: ambi' 'mother;' aruni' 'Dawn;' krshni' 'Night;' 
gandharvi' 'female Gandharva;' gauri' ' cow of the Bos Gaurus;' 
d-Mi' ' messenger ;' nishtigri' ' Indra's mother ;' prapharvi' 
' voluptuous girl ;' meshi' ' ewe ;' yami' ' Yami ;' sinhi' ' lioness ;' 
start' ' farrow cow.' (c) Nineteen others, finally, show no shift of 
accent, nor do they designate animate beings : athari' ' flame ;' 
wit" ' breast ;' ksho'ni' ' flood ;' khdri' ' measure ;' cakrt' ' wheel ;' 
deM' ' dam ;' nadi'' ' stream ;' mahi-nadi ' great-stream ;' nddi' 
'pipe;' ndndi' ' joy -J lakshmi' ' mark ;' vakshi' ' flame ;' Vibdli', 
a river ; vept ' needle ;' sasarpari' ' trumpet ;' sflrrni' ' pipe ;' 
srtrmi' ' flame ;' sphigi', ' hip ;' yayi', ' quick,' is perhaps a root- 
word. Suhdsta might make its feminine suhasti' ; but since 



2-stems.] Noun- Inflection in the Veda. 369 

suhastias, ix.46.4, is m., it is better to read suhastids with BR. It 
is a rule for the Rik that oxytone feminines from barytone 
masculines do not throw the accent forward to the case-ending ; 
and so, in view of sHcrmid, I think it is wrong to set up a stem 
silrmi'. From AV. ii.8.3, paldlid' (pdldla), tilapinjid' (tilapinja), 
however, it would appear that the Rik-rule does not hold for the 
Atharvan (so viliptid's, xii.4.44; hastinid's,v\.'J 0.2; hiranyakepiai, 
v.7.9) ; and that Professor Whitney's suspicion, Ath. Pr. iii.61, 
kalydnyal (vi. L07.3), is well grounded. 

From the Atharvan I add a few words not found in the Rik, 
which belong to C. To 3a belong : Mdi' {M'ta ?) ' fetter ;' 
tandri' (tdndra) ' weariness ;' ttlcshnaprngi' (tikshndprnga) 
'sharp-horned;' pippali' (pippala) 'berry;' saha'sraparni' 
(sahdsraparna), a plant; hastini', vi.70.2, 'female elephant' 
(hastini, ix.3.17, 'having a hand'), cf. p. 368; hiranyakepi' (Mr-) 
' gold-haired.' To 3b belong apvatari' ' she-mule ;' kumar? 
'maid;' viastakepi' and vikepi' 'shaggy hag;' viliptf ? 'cow;' 
vilidhi' 'female monster;' vrkshasarpi' 'tree-serpent.' To 3c 
belong perhaps karkari' ' lute,' a,ndjdtri' ? 'genitrix.' 

Several words in ia have f. stems in i, whose declensional forms 
belong to C : dpia, dpi; samudriya, samudri' (otherwise Gr.) ; 
svaria, svari'. 

About a dozen masculines also belong here : aM, of the same 
meaning as the barytone dhi; tipatht '; nadi' ' caller ;' dakshi ? ; 
prdvi 1 , dushprdvt, suprdvt; yayi'?; ratM, drathi; start', 
sahdsrastan; hiranyavd$i. 

The final radical i is sometimes " split " into iy before the 
vocalic endings of C. This occurs regularly (1) in monosyllabic 
feminine substantives always ; (2) in their adjective compounds 
and (3) those in which the last member is a verbal root, provided 
the I is preceded by a double consonant ; and (4) even after a 
single consonant, if the stem-final is unaccented : thus, (1) piyam, 
dhiy'e, but ddhie; dhiyd'm, but dhind'm, g,rind'm; (2) supriyam, 
but suddhiam; (3) brahmapriyam, but gdihdniam; (4) iithd'- 
dhiye, dirghd'dhiyas, nd'nddhiyas, vd'tapramiyas. Several stems 
other than those included above show the inserted y : yayiyas, 
samiidriyas, and striyas; cf. cakriyau of the Kathaka. This 
seems to me to be a purely graphic peculiarity. ' As is well 
known, it has had a noteworthy extension in the TS. See A. 
Weber, Ind. Stud, xiii.104. 

These three series were presumably once quite distinct — and 
they are so for the most part in the Rigveda; but even here we 
see the process of confusion begun. Now it is plain that either a 
long or a short thematic vowel when standing in the verse before 
another vowel, the initial of the ending, would generally have the 
effect of a short : as, mid', devid'; and furthermore, that when it 
did really become entirely consonantized, the distinction of 
quantity (as, for instance, between the thematic vowels of matya" 
and mahyd') would become lost altogether. The line of separation 
between A and B once becoming thus indistinct, the way was 



370 0. R. Lanman, \i and 

open for the transition of the ^-sterns to the ^-declension. Inasmuch 
as the words belonging to series B are almost exclusively 
feminine, this transition was confined for the most part to the 
feminines in i. Moreover the entire coincidence of certain forms 
of A and B in the Rik would increase this confusion. Thus the 
N.A. and G.L.d. are alike for A and B ; such also is the case with 
the A.p.f. ; and since in B the N.p. is like A.p., the analogy was 
extended to A, and we have a dozen or more nominatives p.f. 
from ^-sterns in -is. So, too, the G.p.f. is alike in each. The later 
language has differentiated some of these cases, or rather utilized 
certain differences by extending them to fields where they did not 
originally belong, and says devyail (like C) for devi', and devyds 
for devi's, N. — forms utterly unknown to the Rik. 

On the other hand, there are a few transitions from the i to the 
^-declension. Thus from ndr is formed the m. and n. adjective 
ndria; but the feminine is from a stem nd'rid, nd'ri. The Rik, 
however, shows forms from a stem nd'ri which cannot be justified 
by the rules of word-formation : they are plain cases of transition 
from B to A. 

This process of transition had already begun in the older Vedic 
time. It would be an interesting and useful task to trace its 
development through the Vedic period. Such a study ought to 
be an eminently historical one, and would, I believe, if carried 
into detail, furnish some valuable clues for the criticism of Vedic 
texts. An example will illustrate this roughly. The Ab.s.f. of 
2-stems ends properly in -es, as ddites. Of these forms there are 
20 in the RV. (from 11 stems); while only three Osteins have 
usurped the endings of series B, namely nd'bhyds, x.90.14, and 
hetyd's, 87.19 — bothhymns of very late date — and bhtt'mids, i.80.4. 
In the AV., on the other hand, there are five ablatives s.f. in -es 
(from 5 stems ; but I may have overlooked some) ; while those in -ids 
or -yds number 21 (from 12 stems). That is to say, the process of 
transition of the s-stems from their own to the 2-declension is 
much farther advanced in the Atharvan than in the Rik. So, on 
the other hand, the transition from B to A had not begun with 
rd'tri in the Rik ; but the AV. shows a number of A-forms (from 
rd'tri). In the Rik, nadt' belongs entirely to C (save one dual 
form, nadi', in a certainly late hymn) ; in this category therefore I 
have had to place it — and that in the face of the grammarians, who 
have used it as a paradigm for B, to which in the later language it 
belongs. In like manner dUti' belongs in the Veda to G, but later 
to B (see BR. s.v.). 

In the case of devdhdti, the transitions in the Rik are confined 
to vocalic cases ; in the Purana we have devah&tim. To follow 
these changes in the post- Vedic writings would lead too far. And 
because in the Vedic period itself everything is so in flux, I have 
thought it best to confine myself to the Rik, in order to get clean 
results. But for convenience of reference I have added some 
material from other texts. Benfey, Vediea, p. 112 et alibi, has 
considered, to be sure en passant, the subject of these transitions. 



2-stems.] Noun- Inflection in the Veda. 371 

His general induction is that the direction of the movement in 
Skt. is from the i to the 1 declension, and opposite to that in the 
Prakrit, where the ?-stems are getting the upper hand. (Cf. Kuhn, 
Pdli-Gr. pp. 79, 81.) For the later Sanhitas and the Brahmanas 
his statement may be right ; hut in the Rik the transitions from 
the \ to the f declension entirely outweigh the others in importance. 
In the following lists I have intended to give with completeness 
for the Rik every stem which has forms other than those belong- 
ing properly to its own declension. 

I. Transitions from A. 

A to B. Twenty-seven i-stems have forms belonging to the 
«-decleijsion. Only four, however, have 2-forms in non-vocalic 
cases : akshi, -i'bhydm (dkshi n. !) ; jdni, N.s. 4; bhtl'mi (has 15 
B-forms), N.s. -t; svd'hdkrti, -ishu. The rest are: dnumati; 
drdti, ,19; avdni; djd'ni; titi; tun; devdhuti; ndkti; nd'bhi; 
nirrti, 3 ; nishkrti; puramdhi, 3 ; pushti; pr'pni, 4 ; bhvji; 
bhrti, 2; yuvati, 2 (see p. 372); vipvdkrshd; $ruti, 2; [preni, see 
N.p.f.] sdmgati; sdyoni; het't. The transition-forms number 69 ; 
that is, the phenomenon is confined to exceedingly narrow limits 
in the Rik, considering that it is a rule of the later language that 
every feminine in i may take the longer B-forms " optionally " in 
the singular. On account of this rule, we need not give from other 
texts instances of transition for non-vocalic cases. 

For consonantal cases, comment and note to TPr. iii.7 give the following from 
the TS., where, however, the pada has? (cf. Vedica, p. 113): from d'hvM, -ibhis, 
ii.6.9 4 ; similarly vyd'hrtibhis, i.6.10 2 : v.5.5 3 ; svd'Mkrttbhyas, vi.3. 9 6 ; hrdduni- 
bhyas, vii.4.13: VS. xxii.26; gacibhis, iv.2.5"- 5 , p. fdci-bhis; if the Rik Pr. made 
a similar statement, we might take all the Rik-forms of this word from a stem 
fd&t; cimhydm, v.7.5 7 ; pronibhydm and prshtM'bhyas, vii.3.16 1 "' 2 ; prshtht'bhis 
AV. xii.1.34, and so also the pada-MS. 

A to C. The instances are few. We have the stem ambi, but 
also from ambi', ambiam; so yayi, but yay'iam, -iyas; sakthi 
(n. !), but sakthi' (f.), -id; sr'ni (I. sr'nid), but also from srni', -id, 
-ias. Karkari of the Rik appears as karkari' in the AV. The 
entirely exceptional art shows the N.s. arts VS. vi.36 ; so BR. 
Cat. Br. sdm ari'r viddm .... sdm prajd' jdnatdm. See 
Z.D.M.G. xxii.576. 

In iv 48.2a, it is unnecessary to pronounce niryuvdno afastias (text -Us): it is a 
pdda of seven syllables; so is viii.39.2d, vlfvd ary6 ardtis; so, too, in both cases 
where Gr. proposes p&ramdhiam, i.l34.3rf and x.64.7a,the pddas are catalectic, 
and we need not change the text (-dhim). 

II. Transitions prom B. 

B to A. About six stems show secondary forms from stems in 
I They are : Sdbhari m., -e 3, -im, -ay as, -indm ? ; Pr'thi m. -im • 

T^L-' 4bhis 2 ; ara nyd™', -k 3, -im; kd'ri, -ibhyas 2, -ishu: 
Oshadhi, -e 2, -is, -im 2, -ayas 3. 

It is hard to say in which category oshadhi properly belonsrs In verse* 
peculiar to the AT. it has the f olio wing forms : (A) -e, 42 ; 2 4 *-L ■ aZ 23 
and vi.96.1 (a false variant of Rik x.91.18 f we must restore tie sZ'rter 



372 G. R. Lanman, [i arid 

grammatical form -is, as the metre shows) ; (B) 4m, 2 ; - id ; -is, N. V. 13 ; -is, A. 
24; 4W»s, 6; -ibhyas, 4; -te<Jm, 23 and vi.15.1, variant of Rik x.97.23; -ishu, 1. 

The same is true of yuvati'. The root yu, yuvdti, would form its participle 
yuvant, feminine yuvaVl. From this stem one half of the Bik-forms may be 
regarded as coming : to wit, yuvatyd's, -yd'm, -V, -yds, -i's A.p. ; but the word 
has lost its participial character so entirely that I have referred it to A . 

The Atharvan has 50 forms from the stem rd'tri (like the Rik); further (A) 
rd'tris, xiii.4.30 ; rd'trim, 16 times, as i.16.1 ; rd'traye, viii.2.20. The AV. has 6 
forms from stem arundhatV ; but arundhate, xix.38.1, which Benfey takes as a 
transition-vocative s.f. (A), may stand for -to (3 pi.). 

The TS. (Weber, Ind. Stud, xiii.105) has: patnayas, 5 times, as v.2.11 2 , q.v. ; 
var&trayas, iv.1.6' 2 : v.1.7 2 ; -trim, iv.2.10 3 ; garbhinayas, ii.1.2 6 ; revatayas, 
v.2.11 1 ; f&kvarayas, v.4.12 2 . 

The Sstman, i.323, has the variant sni'hitim for snehitis, Rik viii.85.13 ; cf. stem 
sni'hiti, i.74.2. Benfey, quoting Siegfried Croldschmidt, adds from the Naigeya- 
Qakha, vii.2.1, rohinlshu (probably a false variant of Rik viii. 82.13, -ishu) and 



B to C. The only undoubted case of transition from B to C is 
seen in the stem stri' ^stitarid') ; it forms striyam, 3 ; -iyas N. ; 
4yas Ace, 2; -tbhis. If we knew that the masculine verbal 
adjective from sparp (sprpdti) were accented prpdna (?), we 
should refer prpanf to C ; as it is, I have referred it to B on 
account of the N.s.f. -?', and regarded -ias (G.s.f.) as a transition- 
form (C). The other transition-forms are prpanias N., yahvias 
Ace, suparnias N., 4as Ace. But it is possible that these may 
be relics of the uncontracted B-forms (p. 366), although in the Rik- 
sanhita these are contracted with almost entire uniformity. I may 
accordingly have erred in assigning khdrt, vakshi', and several 
others to C simply on the evidence of a form that may be merely 
an uncontracted A.s., N. or A.p. It is quite unnecessary to 
pronounce dpipvias in i. 120.8. The pdda is catalectic. 

III. Transitions feom 0. 

C to A. 1. Root-words: dhrayas, ix.54.1 (cf. Qat. Br. xi.3.3 5 , 
dhris, N.s.m.) ; ganapribhis, v.60.86, is certainly a purely metrical 
shortening, and so probably is rtanibhyas, ii.27.12. Benfey, 
Vedica, p. 107 ff., gives send-nibhyas, VS. xvi.26, and pva-mbhyas, 
27 ; and veshapris, TS. iii.5.2 B (see BR.). 

2. The other cases are only sporadic. The accent of onios 
refers oni' to C ; but in i.61.14 we have onim (A), where, 
however, the -im is syllaba anceps. The stem aruni' shows a 
form arundyas in the Pururavas hymn. [Although with some 
hesitation, I have referred aruni' to C on account of the forms 
arunt's N.s., 2 ; -ias, text -i's, G.s. ; -ias Kp. ; -i'ndm, not -ind'm; 
-i'stiu, indifferent. But we have A.p. -i's (B) twice.] Unless now 
we assume that kshoni', like aruni' [and possibly pdkvari (B), 
which has -ayas (A) 'as N.p.f., and -is (C) as N.s.f.], has forms 
that belong in all three categories, I would arrange the article 
kshoni' (cf. Gr.,IF£. 371) as follows: 

G-forms: N.s.f. hshoni's, viii.3.10; so 13.lt, viprd indram kshoni'r avardhayan 
vayd' iva, ' The priests' made Indra grow mighty, as the flood (N.s.f.) the twigs,' 
(A.p.f.)— i. e. as the stream makes the twigs grow ; i.57.4, kshoni'r iva. prdti no 
harya tad vacah, 'Do thou, as doth a flood, (take in) accept our song;' i.54.1, as 



»-stems.] Noun- Inflection in the Veda, 373 

collective and subject of sam drata; i. 173.7, sajoshasa indram mdde [red ?] 
ksJioni'h, ' (In order to get thee into the battle to help us. — thee) Indra, — we are 
united [as] is the crowd (collective) at a feast;' and, finally, as N.s.f., kshont'\s\ 
sacate, i.180.5 (cf. sdsni[s] sarn, x.120.2); kslwnt'bhydm ; 4'bhis, 2. 

A-forms: kshonf, dual, viii.7.22; 88.6: Val. 4.10 (the C-form 
would be -nid); kshondyas, in x.22.9, a verse which is a late 
interpolation in the midst of an old and beautiful hymn. From 
the A V., moreover, we have naptis, Rik napti's. 

C to B. 1. Instead of bhiye we have bhiyd' i-, p. bhiycti, 
viii.64.13. The Rik shows only this one instance of a transition 
which later is so very common with these monosyllabic feminines. 
In i.151.5, takvavi's (Gr., N.p.m.) may be a transition-form (cf. 
siri's), or a N.s.m. as in x.91.2. 

2. The rest are rather doubtful. In i.112.19 and iv.2.16, 
aruni's is A.p.f. In ii.23.14, tapani', I.s.f., also shows a contraction 
peculiar to B. The L.s. of dtiti' ought to be dtitidm. In vi.58.3, 
we have dtltid'm (B). Delbruck, Chrestomathie, p. 79, proposes 
the stem dHtid'= dtitia, taking it as accusative. The B-form, 
nadf, dual, occurs in a late hymn, i.135.9. The un-Vedic form 
nadyo (we need not, with Gr., pronounce -i's) is entirely in 
keeping with the contents of the very late addition, vii.50.4. 

In ix.9.4, pronounce nadio 'jinvad adruhah, and not, with Gr., -i's aj-. Gauri', 
ix.12.3, if it were a dual (Gr., Wb.), would belong here as a B-form ; but it is a 
locative; see Ueb. ii.510. In iii.30. 11, rathi's is better taken as N.s.m. with 
indras (Gr., Wb., A.p.f.) ; in viii.84.1, also, it is just as well to make it a N.s.m: 

IV. Transitions to the N-Declension. 

It seems certain that the forms which here come in question are 
made after the analogy of the m-stems. Thus we could not know, 
without other evidence, whether to divide arcin-d or arci-nd, 
Urmin-d or ■drmp-nd (■drmid), Mrin-d or Mri-nd, Jchddin-d or 
khddi-nd, $vanin-am or pvani-nam, surabhin-d or surabhi-nd 
(cf. atrin-d, dtri-nd, atithin-d, dtithi-nd). But why is the principle 
of analogyso utterly inert in the G.p. ? Here the Osteins lengthen 
the vowel invariably ; the «'»-stems, never (Jcdrin-dm, mdym-dm). 
Cf. Kuhn, Pdli-Gram. p. 80. 

These transitions, if such they should be called, are confined to 
series A, except in the G.p., where the "transition " is universal in 
A, B, and C. In only one single instance, v.44.13, do we have 
dhiyd'm (dhtnd'm occurs seven times); but v.44.1-13 is an 
unintelligible mess of stuff of the latest date. It is not denied 
that dhiyd'm is the more organic form (Schleicher, Cornp.* p. 545 
§ 253) ; but for the assertion that the forms with n are an "indische 
neubildung," the forms nard'm (16; with n, 26), svdsrdm (once; 
svdsrndm, 3), gdvdm (55 ; gondm, 20), and jdguvdm are hardly 
sufficient proof. That the n is lacking here is surely due to the 
half-consonantal character of the stem-final. 

The n is also seen in series A in the I.s.m. (51 forms, 25 stems), 
and f. (! 2 forms, 2 stems), and in the d.n. (hdrint) and p. (-ini, 14 
forms, 4 stems) ; but for all these cases there are examples of the 
more organic form without n. 

vol. x. 51 



374 0. R. Lanman, \i and 

The later language excludes this n from the f., confines it to the 
instrumental in the m.s., and prescribes it for the neuter through- 
out the oblique cases of the s., as also for the vocalic cases of the 
d. and p. The state of things in the Rik is as contrary to this as 
it can be. This is especially clear in the M-stems, where examples 
are more numerous. As stated above, the declension of m., f., and 
n. singular of series A was doubtless once the same, the 
differences being of secondary origin. 

Besides the cases just mentioned, transitions to the w-declension 
are only sporadic or doubtful. Here belong perhaps drminam, 
kirine ?, khddinam, -inas ?, [db/drunam] and pvaninam. 

The following synopses exhibit the terminations of the 
inflectional forms of the several series : 

Series A. Masculine. 

Singular : N., -is, -€, -es; A., -im, -V ; I., -id, -yd, -i ?, -ind; D., -ye, 
-aye; Ab., -es; G., -ias, -yets, -es, -e'/ L., [-ayi, -dy ?] -d, -au; V., -e. 

Dual : N.A.V., -i (iti) ; I.D.Ab., -ibhydm; G.L., -los. 

Plural: N.V., -ayas {-y-as); A., -in, -inr (-y-as, ayas); I., 
-ibhis; D.Ab., -ibhyas; G., -indm, -inaam; -ishu. 

Feminine. 

Singular : N., -is, -t 1 ; A., -im; I., -id, -yd, 4, -i, -ind; D., -aye, 
-i, -I; Ab.G., -es; L., [-ayi, -dy ?] -a, -au, -i; V., -e. 

Dual: N.A.V., 4; I.D.Ab., ibhydm; G.L., -ios, -yos. 

Plural: N.V., -ayas; A., -is (-ayas); I., -ibhis, 4; D.Ab., 
-ibhyas; G., -indm, -inaam; L., -ishu. 

Neuter. 
Singular: N.A., -i; I., 4?; D., -aye; G.,-es; L.,-d. 
Dual : N.A., 4, -int. 
Plural : N.A., 4, -i, 4ni; L., -ishu. 

Series B. Masculine. 
Singular : N., -i; A., im; I., -id; G., -ids, -yds. 
Plural : N., -is. 

Feminine. 

Singular: N., 4; A., -im, (-iam); I., -id, -yd, 4, 4; D., -iai, 
-yai; Ab.G., -ids, -yds; L., -idm, -ydm, 4; V., -i. 

Dual : N.A.V., 4 (iti) ; I.D.Ab., -ibhydm; G.L., -ios, -yos. 

Plural: N.V., -is, -ias; A., 4s; I., -ibhis; D. Ab., -ibhyas; G., 
-indm, -inaam; L., -ishu. 

Series C. Masculine. 

Singular: N., -is, -f; A., -iam; I., -id; D., -ie; G., -ias; N.,4. 

Dual: N.A.V., -id; G., -ios. 

Plural: N.A.,-ms/ 1., -ibhis; G., -indm, -inaam. 

Feminine. 

Singular: N., -is; A., -iam; I., -id; D., -ie; G., -ias; Y., -i. 

Dual : N.A.V., id; I., -ibhydm; G.L., -ios. 

Plural: N.A., -ias; I., -ibhis; D., -ibhyas; &.,4ndm; L., -ishu. 



2-stems.] Noun-Inflection in the Veda. 375 

The following peculiarities are seen in the accent of the 
forms from oxytone stems. It is shifted to the ending, in series 
A, in the I.s., sometimes in the G.s., in the G~.L.d., and as a rule 
in the Gr.p. ; in series B, in the same cases, and also in the D., 
Ab.Gr., and L.s. In 0, on the contrary, the accent never leaves 
the thematic vowel, although it appears in the written text as 
leshaipra svarita (Prat, iii.10). Sporadic exceptions will be 
noticed as they come up. 

Several words have such peculiar irregularities that.it has seemed best to treat 
of them at the end of this section. They are ari, jani, pati, sakhi, and some of 
their compounds. 

Nominative Singulab Masculine. 

A. The ending s is added directly to the stem: as, dvi-s. 
This form occurs 1353 times (from 247 stems). Besides these, 
pdtis occurs 72 times ; arts, 10 ; ndkis, 50, and md'kis, 13 ; the last 
two are indeclinables. Some may be f. : krtis, mushtis, samtanis. 
Frequently occurring forms are : agnts, 389 ; dtithis, 20 ; dhis, 18 ; 
r'shis, 32; kavis, 90; dadis, 15; br'haspdtis, 52 ; brdhmanaspdtis, 
23; rayis, 26 ; vdhnis, 28 ; vicarshanis, 19; p&cis, 38; sdtpatis, 
15; silris, 19; hdris, 58. 

Transitions from C: veshapris, TS. iii.5.2 " : iv.4.1 s : v.3.6 3 ; 
but Qat. Br. viii.5.3 8 (sic), veshaprts; jitatandris, MBh. ; nistan- 
dris, K, Cf. p. 372. 

Besides the regular form vis (6), the stem vi has a form with guna, ves, i.173.1 : 
iii.54.6: vi.3.5: ix.72 5: x.33.2. As parallel to this, in Yacna li.12, Roth takes 
khshnaos of the Copenhagen Codex 4, as N.s.m. of khshrm, -with guna and case- 
ending: noit td im khshnaos, ' Therefore (is no favorer) favors not . . . ' To this I 
would add yaos, N.s.m. of stem yu, ' an ally,' from root yu, ' unite.' Yagna xlvi.18, 
ye maibyd yaos, ahmal (i. e. yo mahyaih *yos, asrnai), ' Whoso is an ally to me, 
him etc' Whoever loves anomalies may translate bodhi dpeh, x.83.6, thus: 'Be 
a friend. ' 

B. See p. 367. Enumeration: ndmi; pr'thi; md'tali; rd'shtri; 
sdbhari, 1 and AV. xviii.3.15. Perhaps we ought to put pvaghni' , 
AV. iv.16.5, here; BR. put it under stem -in. 

C. 1. Root-words. Enumeration: (dhrU, Qat. Br. xi.3.3 B ) 
itthd'-ddhis; kshatrapri's; grdmani's, 2 ; ghrtapri's, 2 ; takvavi's, 
x.91.2 ; darpatapri's; diXrdyddMs; devavt's, 2 ; devdvi's, 6 ; 
patsutah-pi's; padavfs, 6 ; parnavt's; praneni's; madhyaniapt's; 
manyumi's, 2 ; mdryapris; yajiiani's; oapani's; vipvdtodhis; vt's; 
sendnts, 4 ; sionapt's, 2 ; suddhi's, 5 ; (from AV.) prakri's and 
sadyahkri's, xi.7.10; padant's, xi.2.13. Forty-five Rik-forms 
(from 22 stems). 

2. Enumeration: prdvi's; suprdvi's, 4; rathi's, 13, and iii.30.11 
and viii.84.1 (see p. 373); drathis; start's; sahdsrastarts; 
hiranyavdpis. 

Transition from A: art's, VS. vi.36. See p. 371. 

Elision and crasis is actually written in v.7.8, prd svddhittva 
rt'yate, p. -tih-iva, fem. (Rik Pr. iv.13); it is probable in x.51.6, 
rathi'vd'dhvdnam dnu d vartvuh, p. rathi'-iva, better ratM'r-iva; it 



376 C. R. Lanmun, \i and 2-stems. 

is certain in x.84.2 agni'va, text agnir-iva; so in the Atharvan, 7 
times out of 10 (Prat. ii.56). In v.83.3, on the other hand, where 
the text has rathi'va and the padakdra takes it from rathin, 
we must read without elision and crasis rathi'r iva kdpayd' 'pvdn 
abhikshipdn, else the caesura is faulty. 

Nominative Singular Feminine. 

A. The form is like the m. It occurs 470 times (from 136 
stems). Examples are: dditis, 78; nd'bhis, 19; prdmatis, 11; 
bhti'mis, 12; math, 18; yuvatis, 18; rayts, i.66.1 : x.19.3 ; rdtis, 
22; sumatis, 22 ; jdnis, AV. xx.48.2. 

Transitions from B : aranydnis, 3 ; oshadhis. 

Transition from C : naptis, AV. ix.1.3, 10. I know not how to 
explain turydlma yds ta ddipdm drdtis, vi.4.5, except as a fault 
of the tradition, for drdtis. 

B. There is no case-ending. The case-form occurs 731 times 
(from 294 stems). Examples are : yati', 14 ; ghrtd'el, 9 ; jdnitri, 
10; devt, 48; prthivt , 57 ; brhati', 10; maghdni,9; mahi', 35 ; 
uchdnti, 16; sdrasvati, 43; strl', v.61.6: viii.33.19: x.86.6. 

Transitions from A: jdni, iv.52.1 ; b/vd'mi, ix.61.10; bhti'my d' 
dade, p. bhd'mih; but cf. Prat, iv.13 and Orient und Occ. ii.460. 

Transitions from C. I regard kshoni' as standing for kshoni's 
before s-, i.180.5 (see p. 373 top). I am inclined to think that the 
distinction between B and C is already much effaced in the 
Atharvan, in words whose final is not a root ; accordingly we 
have dditrmangali, sumangali' ; nadi', i.8.1: ix.7.14: xii.2.27 ; 
vilipW, xii.4.46, 47. 

Twice Gr. restores mahid, the uncontracted form of mahi' (cf. p. 
356). In i. 167.4, he reads nd rodasi'[ni] dpa nudanta ghorah ?. 

C. 1. Root-words: abhipri's,Z; dhi's,\G; bhi's,2; vratanl's; 
$ris; AV". prakrts, iv.7.6; Vop. iii.80, avis. 

2. Stems in f, not radical (p. 368). We have here 56 forms in 
-i's (from 31 stems) ; and of these 33 occur in the Rik (from 19 
stems). The examples seem accordingly to he numerous enough to 
establish the rule (p. 368 top) that the declension depends on the 
accent. Schleicher, Comp.* p. 510 §246, argues from these forms 
the existence of an ending -s in the N.s.f. of d-stems ; but is it not 
the distinguishing characteristic of most of these j'-stems that 
their final i is not derived from id ? In those whose final certainly 
comes from id (sinhid 1 ), I should rather consider the s as due to 
false analogy (from root-words) than as organic. See p. 355. 

Enumeration : d-durmarigatt's (pdtni), x.85.43 (AV. xiv.2.40, 
-U, see below); aruni's, iv.1.16; 14.3; 'kalydni's {jdyd'), iii. 
53.6; krshni's, vii.7l.l; kshoni's (see p. 372-3), i.54.1 ; 57.4; 
173.7: viii.3.10; 13.17: kshoni'[s] s-, i.180.5, cf. sinhi' rdpdm, p. 
-i'h rA-, TS. vi.2.7 ' ; gandharvi's, x.11.2 ; gaurl's, i.164.41 ; d-titis, 
x.i08.2,3,4: TS. ii.5.11 " ; napti's, ix. 69.3 ; nddi's, x.135.7 ; yami's, 
x.10.9 (AV. xviii.1.10) ; rathi's, v.61.17: 'x.102.2; lakshmi's, 
x.71.2; AV. vii.115.2: xi.7.17: xii.5.6 ; laldmi's, i.100.16 ; vrki's, 
i.117.18; 183.4; pakati's (iva sarjati), x.146.3, 'squeaks like a 



N.s.p.] Noun- Inflection in the Veda. 377 

cart ' (Gr., Wb., A.p.£) ; sasarpari's, iii.53.15,16 ; sumangali's, 
i.113.12: x.85.33; [AV., -W, iii.10.2 : xiv.l. 60 ; 2.25,26 ; but not 
" der Regel gemass " ( Vedica, p. 1 1 6), since the change of accent 
requires inflection according to C. These " regularities " of the 
later Vedic texts are rather an efiacement of distinctions which 
prevail in the Rik almost without exception. In the Rik-passage, 
xiv.2.28, accordingly, we have -It's.'] start's, i.122.2: vii.101.3: 
x.31.10: VS. iii.34 : *TS.i.5.6 * : Kath. vii.l ;— from the Atharvan: 
jdtri's, xx.48.2 (MSS. ; ed. jdnis); tandrt's, viii.8.9 : xi.8.19; 
[MBh., atandris, gatatandris, jitatandrts ; R., vyapanitatandri;] 
ydtudhdni's, i.28.4; — from the TS. (given by A. Weber, Ind. 
Stud, xiii.101) : mahishi's !, i.2.12 2 (mdhishi, Rik, from mahishd) ; 
vipvartipi's, i.5.6 \8 '' ; pabali's, iv.3.11 5 ; sinhi's, i.2. 12*. 

Transitions from A : drdtis, see p. 376 ; in arm pyeni sacate 
vartant'r aha, i.140.9, p. -nih, we have a purely metrical lengthen- 
ing; Prat, iv.13: 'A black path follows along after him.' Vedica, 
p. 116. Gr. as A.p.f. 

Transitions from B: pdkvaris, TS. hi. 4.4 ', looks to me 
suspicious (or may it not be N.p.f. ?) ; still more so is sahd devi'r 
arundhati', AV. vi.59.2. 

Elision and crasis: see p. 375. 

Nominative and Accusative Singular Neuter. 

A. The neuter has no case-ending. Kim, which is gradually 
supplanting the more antique kdd even in the Veda, is perhaps 
a phonetic variation of *kd-m. It occurs 49 times ; and as adv., 
31 ; cf. ndkim and md'Mm. The case-form occurs 256 times in 
the Rik (from 37 stems) ; all but 87 of these belong to bhti'ri, 
mdhi, and svasti. The list is : dkshi; dkshiti, 3 ; agavyllti; 
djdmi, 3 ; anji, 5 ; aprati, 2 ; abhibhtiti, 3 ; abhimdti; dsdmi, 9 ; 
itdsUti, 3 ; ghdshi; codaydnmati; jd'grvi; jdmi, 4 ; tardni; 
tuvishvdni; trind'bhi; dharnasl; nidhmvi; pdpuri; parogavyMi, 
as adv. ; bhurvdni; bhil'ri,47 ; mdhi, 84; vii.30.1 (Gr., mah-'e) ; 
iv.56.5 and x.93.1 (Gr., mdhi, V.d.f.); vipvdcarshani; vr'shni, 3; 
$dmtdti; pdmi, viii.45.27, BR.; otherwise Gr. ; piici, 11 ; sdkthi, 
2; sdnerni, 2, and as adv., 6; [sdsni : sum, x.120.2 (AV. v.2.2); 
Gr. reads sdsnis : sdm; cf. kshoni' s-, p. 373] sdnasi, 2 ; surabhi; 
stM'ri; \svdni, vi.46.14, BR. and Gr., Ueb., take as a verb;] 
suabhishti; suasti, 33, and svasti, 2 ; hd'rdi, 9. From the 
Atharvan : dkshi, 3 ; dsthi, 8 ; ekanemi, 2 ; dddhi; pr'pni, i.11.4 ; 
bhd'ri, 4 and v.2.3 (Rik, vipve) ; mdhi, 6 ; vddhri, iii.9^2 ; surabhi; 
sakthi, xx. 136.5 (MSS. saktnm). 

C. The only place in the whole Rik, to my knowledge, where a 
N.A.s.n. of a stem in * radical is needed is Val. 2.10, ydthd 
gdparye dsishdso adrivo : mdyi gotrdm haripriyam (sc. sishdsa) : 
'As thou wast ready to give to G., so (give) to me a fair tawny 
herd.' That is, of the form prescribed by the grammarians (as 
jalapi), there is not a single example. We should expect haripri; 
but rather than use so unfamiliar a form, the poet puts a masculine 
adjective with the neuter gotrdm. Cf p. 343 and V.s.m. 



378 G. R. Lanman, [i and t-stems. 

Accusative Singular Masculine. 

A. The ending -m is added directly to the stem. The form 
occurs 1208 times (from 205 stems). Besides these, arim occurs 
twice ; patim, 49 times. Examples are : agriim, 269 ; dtithim, 25 ; 
ddrim., 30 ; dhirn, 40 ; tirmim, 23 ; kavim, 28 ; ptiramdhirn, 6 and 
i.134.3: x.64.7 ; br haspdtim, 25 ; yayim, 2 ; ydnim, 61 ; rayim, 
1 80 ; hdrim, 24. 

Transitions from B : pr'thim; sdbharim. 

Transition from C : onirn, or f. 

Transition to w-declension : hhddinam ?, vi. 16.40; 6! yarn, haste 
nd hhddinam p'ipum jdtdm nd b'ibhrati, ' whom they carry as a 
ring on the hand, as a new-born child ' (on the arm) ; so BR. 
Here I would put iirmi-nam (Urminam : fXrrnim :: ■drmind : ttrmid'), 
ix.98.6, and pUlapdni-nam, MBh. ; but not abhimdtin-am, i.85.3, 
as does Benfey, Ve'dica, p. 124; for the accent would be wrong, 
and Lindner, p. 124, gives six similar formations. See stems in in. 

B. Here belongs ndmim, vi.20.6. 

C. 1. Root-words. (a) gdthdniam, viii.81.2; durddhiam; 
devdviam, 3; pratlviam, 3; yajnaniam; suddhiam. (b) Cf. p. 369. 
abhipriyam; brahmapriyam, 2; ghrtapriyarn, AV. xii.1.20; 
-priyam at the end of compounds : adhvara-, 2 ; kshatra-; ghrtd-; 
jana-; yajna-; su-, 2; hari-, viii.15.4. 

2. Stems in i, not radical: rathiam, 8; mprdviam. 

Transitions from A : yayiam, ii.37.5 ; it is needless to read, as 
Gr. proposes twice, ptirarhdhiam, text-im. See p. 371. 

Transition to w-declension. Benfey, Vedica, p. 122, sees an 
instance of this in pvan/imarn, VS. xxx.7 ; this he refers to pva-nt ; 
BR, to pvanin; Mahtdhara : puno netdram. Cf. Kuhn, Pdli- 
Gram. p. 81. 

Elision and crasis: AV. agni'va, text agnim-wa, viii.2.4 (so 3 
times out of 5, Prat, ii.56). 

Accusative Singular Feminine. 

A. The form is like that of the masculine. It occurs 607 times 
(from 156 stems). Examples are: dditim, 32; ptiramdhim, 5 ; 
bhil'mim, 19; matim, 22; rayim, iv.34.2: v.33.6: vi.8.5: x.167.1; 
rdtim,20; vrshtim, 26 ; surnatim, 41 ; mshtutlm, 35; sdmrtim, 
or m. 

Transitions from B : aranydnim; onirn, or m. ; Oshadhim; 
sni'Mtim; SV. i.323. 

B. The form occurs 303 times (from 116 stems). Examples 
are: urvi'm, 9; tdvishim, 13; devi'm,)8; pipy&shim, 9; prthivi'm, 
62; in i.67.5 : vii.99.3, pronounce prthvtm; mahi'm, 35 ; for -Mm, 
x.50.5, see p. 357. In ii.11.8, is an example of the uncontracted 
form : d&re pdr'e vd'niam vardhdyanta, text 4m. 

Transition from C': nadi'm only, from AV. iv.37.3. 

C. 1. Root-words. (a) devdviam (tvdcam), ix.74.5.^ (b) 
abhipriyam, 2; dhiyam, 45; dhiyamdhiyam, 2; bhiyam; 
priyam, 14. 

Transition from B : striyam, 3 ; so AV. viii.6.9,16,25. 



A.s.f.] Noun- Inflection in the Veda. 379 

2. We have here 18 Rik-forms (from 16 stems). Enumera- 
tion : atharvlam; ardyiam; gauriam, 2 ; nadiam; napt'iam; 
ndndtam; prapharv'iam • yamiam; vibdltam; vipvarupiam; 
vrh'iarn; sinhlam; sHrmtam/ stariam, 2 ; sphigiam; svariam, 
x. 177.3 (cf. accent of svari'ndm). Even in the Atharvan the 
thematic final is rarely, if ever, consonantized : thus, kumdriam, 
xiv.1.63 ; ktidiam, v. \9.12; lakshmiam, i.18.1 ; laldmiam, i.18.1, 
4; viliptiam, xii.4.41 ; vilidh'iam, i. 18.4; so napt'iam, i.28.4 ; 
prapharviam, iii.17.3: v.22.7. 

Transition from A : amb'tam. 

The only place in the whole Rik — except vii.50.4 (p. 373) — where the final V of 
a word belonging to series C is consonantized is vii.68.8<2, starydm. We cannot 
read stariam cic chaktydfvind facibhih; see p. 380. The preceding seven stanzas 
are each of three verses, and the metre shows plainly that the fourth verse here is 
a later addition; while in verse 9, d is the usual refrain. Gr. (Ueb.) brackets 
them both. 

Instrumental Singular Masculine. 

A. The general ending is added directly to the stem : 1. 
■drmid' (accent), i.184.2; pdtid, x.85.22 ; sdkhid, vi.56.2 : viii.48.10: 
x.6.2; 71.10. 

2. pavyd', i.88.2: v.52.9: vi.8.5; rayyd', x.19.7; pdtyd, 
x.85.24,27,36; sdkhyd, i.53.7 : vi.21.7 : viii.43.14 : x.50;2. So in 
the AV. : pdtyd, ii.36.1,4 : vi.22.3: x.1.3 : xi.1.14: xiv.l. 40,52; 
2.32; rayyd', iii.14.1: vi.78.2. 

3. The termination -id is often contracted to -i in the feminine. 
The only example of this in the masculine is ghr'ni, ii.33.6, 
ghr'niva chdyd'm arapd' afiya, p. -ni-iva, ' As by the heat 
unharmed, to shelter bring me ;' but it may be Ab. with elision 
and crasis (cf. vi. 16.38). Otherwise Weber, Ind. Stud, xiii.58. 

Transition to the w-declension. This occurs in 51 forms (from 
25 stems) : ugrund, 7 ; ddrind, 2 ; as'ind; dhind, 6 ; ■drmmd, 7 ; 
kavind, 2 ; kdfind, 2 ; kikidw'ind ; jamddagnind, 4 ; devd'pind; 
dhds'md; panind, 2; pdtind, iv.57.1; pdprind; br 'haspdtind, 2; 
man'ind; yaymd; ray'ond; rapmind; vddhrind; vavrind; 
wr'shandbhind; vrshnma; pucind, ii.5.4 ; 35.8 (or n.) ; sdsnind. 
AV. has -pdtind, xvi.6.6 bis. 

B. Here belongs only ndmid, i.53.7. 

C. 1. Root-words: (a) dildhid,ix.5S.S ; (b) vdmanena kavipriyd, 
Vamana, Introduction. 2. rathid, iii.36.6: vii.95. 1. 

Instrumental Singular Feminine. 
A. i. Here belong 37 forms (from 19 stems): (a) with shift of 
accent, i. e. oxytone stems : Utid', 4, and viii.18.7, coalescing with 
d'; pitid' ; vrshtid', 2 ; sukirtid' ; sumatid' ' ; sushtutid' [and from 
the AV., pushtid', xix.31.2,9,13 ; in x.6.27, Mrtid is a palpable 
error for -tid! ' ; cf. C] : — (b) barytone stems : dvartid; d'bhUtid; 
jdlpid; tv'tshid; dhrd'jid; pd'rshnid m. ? ; p-dramdMd, 9 ; bhU'mid, 
i.161.14; rdnhid, 5 ; rdjid; vi'cid; sr'nid; hdrshid, 2 [and 
from the AV, abh't$astid, xii.5.58; drdtid, iii.31.1 ; tr'ptid, 
ix.5.9; tvlshid, x.6.27 ; dhrd'jid, iii.1.5 ; 2.3; pd'rshnid, vi.42.3 : 
viii.6.17; bhU/tid, x.6.24 ; stibhUtid, iii.14.1 ; rucid, xiii.2.30]. 



380 0. R. Lanman, [i and 2-stems. 

2. Here belong 13 forms (from 11 stems): (a) oxytones : 
apityd' ; matyd' ; mithatyd! ; vasatyd' ; sumatyd', 3 ; mshtutyd! ', 
making hiatus with dl: — (b) apdnyd; d'/cdtyd; tshtyd; devdhH- 
tyd; prdmatyd. AV. xii.3.33, svddhityd. 

In the Rik the forms in -id are to those in -yd about as 3 : 1 ; in 
the Atharvan (according to my notes; I fear they are not com- 
plete), about as 1:5; that is, the relations are most strikingly 
reversed. The Atharva variant, dcittyd c'ed, vi.51.3, compared 
with dcitti ydd of the Rik, vii.89.5, betrays its later character as 
well by the case-form as by the word c'ed. 

3. The termination -id is contracted to 4. Forms of this kind, 
where both texts have 4, occur in the Rik 105 times (from 35 
stems), and are therefore more than twice as frequent as those in 
-id and -yd put together (50). They stand frequently at the end 
of a pdda (cf. Mi', matt'). They were recognized by the native 
authorities: Schol. to Pan. vii.1.39, sushtuti', sushtutyd iti loke. 

(a) The pada has always 4. In i.20.4c; 164.85: iv.4.145: 
viii.23.14a/ 76. 6a": ix.9.2c: Val. 3.3c end, the 4 stands before a- 
or d-, is written y, and is to be pronounced with hiatus ; in 
vii.68.8a", accordingly, we must also pronounce pakti' ap- (see p. 
379). Only twice is the hiatus written, in vi.5.7 (end of pdda !), 
and x.64.1. It is written as coalescing with a following i- or i-, 
and is so to be pronounced in ii.5.6a*/ 13.2a : x.20.66 (trochaic) ; 
but with hiatus in vii.28.3a. Enumeration : dcitti, 2 ; dprabh'Mi; 
avyathi ' ; d'huti; tttt ', 26 and x.64.1 ; rjuniti'; citti; jushti; 
di'dhiti; dushtuti' ; dushtuti; dhiti', i. 164.8 : v. 25.3 with swmatid' ; 
nipiti; pdrivishti; purdjiti; prdniti, 4 andiii.51.7 (L. ?) ; prdbhMi; 
prdyati; prdyuii; metti', 15; vartani'; vishti'; viti' 3, and ix.61.1 
(ayd' viti'): vi. 16.46; vrshti', 2; pakti'; pdkti; grushti', 13; 
sa'dhri', ii.13.2, I take as Ls.f. of a stem sadhri; or from sddhri 
(Lindner, p. 102) with adverbial shift of accent ; ' with an aim,' i. e. 
intently ; sdh-Ati, 2 ; suditi'; stiniti, 2 ; surnati' , 2 ; sdmiti; 
sushtuti', 5 ; hdstacyuti. In the AV. we have dhiti', vii.1.1 ; and 
several forms in Rik-passages, vii.48.1 : ix.9.8: xviii.1.23. In Rik 
vi. 16.46, we may, with Gr., restore the longer grammatical form 
vitid', text viti ; but the caesura is better if we read viti' . . 
mdrtio (text -to). 

(b) The pada has 4: i.182.5, supaptani'. Prat, vii.15. 

(c) The samhitd has 4 before r-, and the pada reads -ih: 
vi.18.10, .... indra heti': rdksho m dhakshi, Gr., Ueb. i.578, 
'Burn up (destroy) the evil one with thy bolt;' BR. suggest 
indrahetih; i. 180.4, tdd vdm . . pdpva4shti : rdthyeva cakrd! 
prdti yanti mddhvah, ' Therefore (because ye blest Atri, according 
to his wish, eshe) to you, with desire for kine, (swift) as chariot- 
wheels, our oblations go ' (and ye ought to do the same by us). 
Contrariwise, vi.13.1, prushti' r-, p. 4', Gr. reads 4h (Ueb.) — 
wrongly, I think. 

4. The final 4 becomes 4. Cf. uta U dfrindni dfriti (Copen- 
hagen 10 -ti), ' uta tvdm dprindni *dpriti (dpribhis),'' Vd. xxii.5 ; so 
dyece y&ti, Ye. ii.l. This happens in all but three instances at 



I.s.f.] Noun- Inflection in the Veda. 381 

the end of the pdda. It occurs 25 times : viz., (a) at the end of 
pdda b or of theverse : ishdni; prdyukti, 2; vdshatkrti; sadhdstuti; 
suvrkti, 4; supasti; suasti, 4; hav'tshkrti: — (b) at the end of a 
pdda before vowels: ■dpapruti, 2; upd'bhrti; suvrkti: in-, i.61Ac, 
16a/ supasti: {ipa, vi.67.3; here Gr., Ueb. i.524, would place rtti, 
ii.39.5, text ritir a-: — (c) in the interior of a pdda: trivishti 
(eti), iv.6.4 ; (adh-) 15.2; nitikti (yd), vi.4.5 ; here I would place 
dsmrti, ' (with, i. e.) out of forgetfulness,' AV. vii.106.1. Gr. puts 
here ablushti (pdasi), ii.20.2, p. abhishti-pd' asi jdndn. Weber 
sees here crasis after elision (as in iv.46.le), -pd'h. See Ind. Stud. 
xiii.58,104, and Whitney to TPr. x.13. 

Transitions to the w-declension : dhdsind, vi.67.6; nd'bhind, 
vi.39.4. In AV. vi.141.2, svddhitind is m., although svddhityd 
occurs xii.3.33. 

B. 1. Here belong 49 forms (from 27 stems), (a) Oxytones: 
asiknid' (? p. 368) ; devid', 5 ; purdnid' ; samdnid', 5 ; sHcid': — 
(b) dnvid, 3 ; dpvdvatid; Jcundrnd'cid; ghrtd'cid; citdntid; 
citdyantid; tmdnid; devd'cid; ddvidyutatid ; ndvyasid; pdtnid, 
2; rdhinid; vdsvid; vipvd'cid; vipvid, adv. ; $dcid,\\; pdmid, 
2 ; parmaydntid; pimid; satrd'cid; stdbhantid; hdrinid. Add 
from AV. : apvdbhidhd'nid, iv.36.10: v.14.6 ; asiknid, v.13.8; 
vd'sid, x.6.3 ; sautrdmanid' , iii.3.2. 

2. There are 22 forms (from 12 stems). (a) Oxytones: 
prthivyd', 8 ; mahyd' ; sddhdranyd' (? p. 368) : — (b) dpvdvatyd; 
gdtumdtyd; tmdnyd; ndvyasyd; mddhumatyd; vd'javatyd; 
fdcyd, iv.35.5 ter (once with hiatus and once with fusion before a-) ; 
sdmavatyd; hiranydyd (for hiranydyy-d), viii.1.32 ; 67.2. Prom 
AV. : annddyd', xv.14.4,5; daivyd, viii.1.3; vipvdbheshqjid, 
vi.136.3; sdrasvatyd, 2. TS. and VS. have urvyd'; but Rik, 
urviyd', 23. See BR. 

3." Qdmi, i.110.4: iv.22.8: ix.74.7: and probably x.92. 12, text 
paminahushi; supdmt, vii.16.2 : x.28.12. 

4. (a) At the end of a, pdda: pdmi, ii.31.6 : viii.45.27: x.40.1; 
supdmi, v.87.9a/ Gr. would convert the cadence from trishtubh to' 
jagati by restoring the longer grammatical form supdmid. ' (c) In 
the interior of a pdda: pdmi r'-, i.87.5; pami d-, iii.55.3. Here 
we may put urvi- of the compound urvi-titih, vi.24.2. We have 
supdmi, VS. i.15 : TS. i.1.5 3 : TBr. iii.6.6 4 . 

C. 1. Root-words: (a) ddhid', AV. vi. 132. 1-5, accent! (b) 
dhiyd', 75; priyd' ', 15; bhiyd', 14; avadyabhiyd', x.107.3, 
' through fear of blame :' it is accented as though the first member 
of the tatpurusha were an independent word. The hymn is very 
late. 

2. (a) mandtikid; vepid; stirmid, vii.1.3 ; sphig'td. Srntd, 
x. 106.6, is dual.' (b) cakriyd, ii. 3 4. 14 (cf. N.A.d.f., C). For 'the 
accent of Ulapinjid' and paldlid! , AV. ii.8.3, and of sahasraparnid' , 
vi.139.1, see p. 369. In x.6.27, MSS. and ed. have kirtld, which 
must be an error for kirtid! . 

Tapani', ii.23.14, if it stands for tapanid, belongs here; other- 
wise, under B. 

vol. x. 52 



382 0. R. Lanman, [i and ^-sterns. 

Instrumental Singular Neuter. 

A. B. C. There is no certain example. In ii.35.8, pucind may 
be masculine ; for in iv.2.16, p&ci is rather adverb (Gr., Ueb.) than 
substantive. In v.42.18, Gr. takes suprdniti as adj. with dvasd; 
BR., as fem. substantive. 

Dative Singular Masculine. 

A. 1. The direct combination of simple stem and ending is seen 
only in pdtye (8) and sdkhye (10). In no case does the metre 
show -ie. 

2. The usual form is made by gunating the stem-final and adding 
the general ending -e: thus, agn'e-e, agndye. This form occurs 
148 times (from 44 stems). In kshetrasya pdtaye, AV. ii.8.5, and 
bhdtdsya pdtaye, iii.10.9,10, the second word is declined as if it 
were part of a compound. This it is virtually; cf. bMtapdtis, 
ii.14.4. In brhddrdye, Rik i.57.1, we have, I think, a clear case of 
metrical shortening of the penultimate of ajagati-pdda. It stands 
for brhddrdye. It is precisely so with rdhddrtiyas, viii.46.23 ; the 
& is penultimate of a gdyatri-pdda. In each case we should set 
up the stem -rat, not -ri or -ri. 

Datives s.m. in -aye are found from the following stems : agni, 48 ; dtrri, 12 : aditi ; 
arati; art; afvamishti ; dhi, 3; dpi, 3; r'shi, 3; kavi, 8 ; kiri, 2; kukshi; 
krandadishti ; gdvishti, 3 ; ghr'shvi, 6; cakri; jagmi,2; jasuri,2; jdmi; t&ji; 
tuvigri; twrvi'ti,i\ tu'rvi; dabht'ti, 5 ; dafoni; d&dhi; prabMU; Ir'haspati ; 
bhandddishti ; bM'rni, 2 ; yati; vahni; fuel ; sani, 12; saptavadhri, 2 ; sd'rathi, 
2; sukhddi; sudtti; sum&jjdni; sushvi,i; suri,2; syu'marafmi; Jidri; havydddti. 

C. 1 . Root- words : (a) jihmapie; durddMe; dUd/ue : — (b) 
itthd'dhiye, 2 ; yajnapriye; and VS. xxii.30, ganapr'tye. Cf. kiye- 
in kiyedhd's with md'-M-m. 

2. Here belong ratMe, viii.44.2'7, and suprdv'te, x.125.2. In the 
Atharvan variant of the latter verse, we read suprdv'id, iv.30.6 ! 
See note to Ath. Prat, iv.ll. 

Dative Singular Feminine. 
A. 1. In no instance is the ending combined with the simple 
stem. 

2. The usual form is like the m. It occurs 507 times (from 50 
stems). The average number of occurrences of each stem is here 
verv unusually large. It is used with especial frequency as an 
infinitive : thus, ishtdye, pitdye. 

Enumeration. Datives s.f. in -aye are found from the stems : ajiti; aditi; dditi 
5; abhim&ti; abhifasti, 3 ; abhislai, 16; dmati,1\ arishtdtdti, 4 , ahati; dbhogi 
ishi- ishti, 20; uti. 88; gdvishti: g&rti; grbMtdtdti; luji; durbhrti; devdtdti, 10 
devaviti, 22; dhiti; nrti; nr'piti; pakli; piti, 67; puramdhi ; purvdcitti, 8 
p&rvapiti, 4; prdttti; prdtdrti; prafaxti, 9 ; bhakti; bharahuti, 2 ; maghdtti, 6 
mitrddhiti; medhdsdti, 5; rdti; vdsutti,2; vdsyaistiti, A; vd'jasdti, 34; viti, 31 
crusbti ; sdrvatdti, 4 ; sdti, 34 ; sikti ; sukshiti ; sumati, 2 ; sdmapiti, 49 ; suasti, 
45 ; svasti ; havydddti, 6. 

3. Several stems appear to have a dative in 4. It is of unclear 
origin. They are: dnUtt, vi.29.6 ; Uti', i.100.1-15 ; 178.1: iv.23.2; 
25 2: 34.9: vi.24.9 ; 29.6: vii.25.1 ; 59.9 (AV. vii.77.1), 10; 



D.s.f.] Noun- Inflection in the Veda. 383 

viii.21.7; 57.4; 86. 7; 88.7: ix.97.38: x.15.4 (AV. xviii.1.51); 
35.13; 104.4; vUi', ix.91.2; 97.49; vrshtl', v.53.5, may be I.a.f. 
(p. 380) ; and so prdniti, vii.28.3 {Orient und Occ. ii.466). 

4. In v.51.12, we have saastdye thrice and suasti, evidently in 
the same construction, once; the same is true inverse 13. In 
viii.Sl.ll, suasti occurs within pdda b before s-, and in c we have 
suastdye. In v.42.15 it occurs at the end of c before u-, and at 
the end of d in i. 174.9 = vi. 20.1 2: iv.11.6: v.4.11. 

B. The general ending -e was added to the stem : thus, devid-e, 
deoiai, devyai. To the apprehension of the speaker eventually, 
the parts were doubtless devy-ai. There are 21 forms (from 13 
stems): 1. vippdtniai: — 2. (a) devyai; prthivyai, 6; brhat- 
yai; mahyai, 4 ; sinivdlyai; (b) dviatyai; 'vyatyai; jit'ryantyai; 
jyd'yasyai; dhenumdtyai; papumdtyai ; vd'javatyai. From the 

AV. I have noted 21 forms (from 14 stems), not counting two in 
vii.46.2. Under 1. fall Mranyakepiai, v. 7.9, and striyai, v.30.3 ; 
parjdnyapatniai, xii.1.42, and sahdpatniai, xiv.1.58 ; the rest, 
under 2. In vi.107.3, kalydnyal is doubtless an error; read -ya% 
(p. 369). 

Transitions from A: 1. devdhtitiai, viii.39.4; prutiai, ii.2.7: 
x.111.3: — 2. in the tenth mandala, pushtyal, 105.5; turya'i, 
pushtyal, and bhujyai, 106.4; nlrrtyai, 164.1; bhrtyafi, 29.4. 
From the Atharvan I have noted at least 13 transition-forms in 
-yai (from 11 stems). In vi.41.15, the metre requires us to restore 
the secondary and shorter grammatical form, d'lcMyai (utd 
cittaye) ; text d'JcAtaye. 

Transition from C: bhiyai, viii. 64.13. 

C. 1. Root-words: (a) ddhie, x.95.13 :— (b) dhiye, 6; priy'e, 34. 
2. Here belong ndndie (with mude), i. 145.4; tueshte (with 

meshd'ya), i.43.6; vrkie, i.116.16 ; 117.17: vi.51.6. 

Dative Singular Neuter. 

A.B.C. The only example known to me is pticaye (padd'ya), 
vii.41.6 (AV. iii.16.6). 

Ablative Singular Masculine. 

A. 1. Of a form in i-as there is no example. 

2. Of the usual form there are 14 occurrences (from 8 stems) : 
ddres, 2; dhes; udadhes; gir'es, 5; ghr'nes, vi. 16.38 (of. p. 379) ; 
ydnes, 2 ; patdmiltes; sdmrtes. From' the AV. : tirapcirdjes, 
vii.56.1 ; parnadhes, iv.6.5 ; pldpes, ii.33.4. For girdyas (Ab. ?), 
see N. p. m. • There are no examples for B and C. 

Ablative Singular Feminine. 

A. 1. At the end of the catalectic pddas i.l28.7/and g, of 11 
and 7 syllables respectively, Gr. proposes to read dhUrtias; but 
the metre is good as it stands (dh-iirtes). 

^ 2. Here belong 20 forms (from 11 stems). They are: ddites; 
dpttes; abhlpastes, 8; abhihrutes; dmates; drdtes; dhdses; 
dhfirtes, 2 ; pdrishHtes; vasates, 2 ; srutes. From AV. abhlpastes, 



384 C. R. Lanman, [i and 2-stems. 

vii.53.1 ; d'hutes, iii.22.4; nrt'es ?, vi.18.3; bhU'mes, xiii.4.35 ; 
vasatbs, vi.83.1. 

B. 1. Here belong prthivid's, ix.8.8 ; 31.2; 57.4, and ji'vantids, 
v.78.9. In every instance the wor-d occupies places 3-6 in a pdda 
of 8, and is followed by ddhi. 

2. Prthivyd's, 18; mahyd's; urvdfyds; pdtantyds. The AV. 
has prthivyd's, 19; brhatyd's, viii.9.4 ; avadydvatyds, vii.103.1. 

Transitions from A: bk&'rnids, i.80.4; hetyd's, x.87.19; nd'bhyds, 
90.14. 

From the Atharvan we have : Tietid's, iv.10.5; dbhUtids, vii.100.1 ; drdtids, x.3.7 : 
xiii.4.47; dcastids, xii.2.12; d'hutids, xii.1.13 ; nd'bhids. ii.33.4: ix.8.12 ; bh&'mids, 
ii.30.1 : — hrshyd's, ii.4.6 ; devahetyd's, viii.1.12 ; cirshahtyd's, i.12.3 ; bh&'rnyds, 
iv.19.6: x.1.13: xii.3.26: xix.16.2 ; 28.4; rd'tryds, xiii.4.30. In vi.121.4, we 
have elision and craais : y&nyeva prdcyuto gdrbhah, text, -yd iva. 

C. There is no example of an Ab.s. 

Ablative Singular Neuter. 
A.B.C. For the neuter there is no example. 

Genitive Singular Masculine. 

A. 1. Here belongs ari-ds, in vii.8.1 ; but not necessarily in 
iv.48.16 and vi.l4.3&y furthermore, aryds, ii.23.15 and 34 times 
(in x.27.8 bis, it is N.p.f.); dvyas, 17 times and ix.108.5. The 
resolution sddhrias, at the end of v. 44. 106, makes the pdda one of 
12 (text, -es, 11) syllables. 

2. The form in -es occurs 162 times (from 42 stems). 

Enumeration: agues, 55; aires, 4; ddites, vii.82.10; ddres, 10; dsushves; ahes, 
3; djfo, 2; apes, 3; ishvdhes; Urmes; r'shes, 2; kales; haves, 4; hires, i.31.13: 
ii.12.6: x.41.2 ; gires, i.56.3 ; gaurivttes ; ghr'shves, 3; jivres; dadrcdndpaves ; 
dabhi'tes ; dudhes; dr'tes ; dhdses, 4; namuces, 4; panes, 8; pates, ix.35.6; 
plate's; br'haspdtes, 6; brdhmanaspdtes, 2; bh&'res, 4; milrdtithes; medhidtithes ; 
yayd'tes ; vavres, 2 ; vdhnes ; ves, 14 ; vrshd'kapes, 2 ; sdptes ; sd'varnes ; sushves ; 
sHres, 5 ; hares. 

B. Here belong Tirapcid's, Pr'thyds, and Sdbharyds. 

C. 1. Root-words: (a) gdthdmas, i. 190.1; dddhias, 2 ; manyu- 
mias; sudhias : — (b) gana$r'iyas. 

2. Here belong uhias, x.144.4, and suprdvias. 

Transition from B : The stem of ahighnyds (drvatas), AV. x.4.7, 
must be ahighni' (properly a feminine formation like dpatighnt) ; 
but I know of no better place than this in which to put it. The 
accent is peculiar to the Atharvan (p. 369 top). 

As mentioned on p. 382, the 8, of rdhddrayas, viii.46.23, is purely metrical. 
Otherwise wo must assume that the fuller ending -as has been added to the 
gunated form, -re, of stem -ri (BR.). If any similar instances — other than 
citrotayas, x.140.3 (A.p.f., Roth), vayas, i.104.1 (A.p.m.), and facayas, AV. v.1.3 
(A.p.f.) — can he found, they are merely sporadic. See stems in -<»», G.s.m. 

Genitive Singular Feminine. 
A. 2. Of the usual form there are 44 occurrences (from 11 
stems). Enumeration : ddites, 23 ; abhipastes; dmates, 2 ; ishtes; 
krshtes; dendrites; nirrtes, 6 ; pushtbs, iv.41.10 : v.41.20 (m. ?) ; 
pr'pies, 4; vrshtes, ix.41.3; 74.3; vratates. I have noted 16 
forms (from 8 stems) in verses peculiar to the AV. 



Gr.s.f,] Noun- Inflection in the Veda. 385 

B. Of the 99 Rik-forms (from 19 stems) only 15 end in -ids; the 
rest, 84, in -yds. Enumeration: 1. yatid's; urvid's; devid's; 
prthivid's, 7 ; ydtid's; stri-y-d's; — upindrdnids; rd'trids; 
sil 'shyantids : — 2. prthivyd's, 68; tnahatyd's; vadhrimatyd's, 
5 ; — anpumdtyds, 2 ; Hrjdyantyds; tdvi.shyds; dd'numatyds; 
mdnspdcanyds ; rd'tryds; vivdsvatyds; pdmyds; $6pncatyds. 

From the Atharvan I have noted: punyd'f ca caturakshid's, iv.20.T; devid's, 
xiv.1.63: pipdcid's, i.16.3; prthivid's, vi.86.1,2: xviii.2.20; vefantid's, i.3.7 ; 
ratharvid's, x.4.5; stri-y-d's, v.11.8: xii.2.39; hastinid's, vi.70.2; — amucyds, 
xvi.6.10 ; grhapatnids, iii.24.6; patnids, ix.3.5: — and, besides, 64 forms (some Ab.) 
in -yds (from 18 stems). 

Transitions from A : 1. drdtids, ix.79.3 bis; bh&'mids, p. -d, 
x.75.3, with elision and orasis, bhO! miopdri : — 2. yuvatyd's; 
dnumatyds; nirrtyds, x.165.1, a late hymn — 'The Death-dove;' 
pr'pnyds, 4 ; bhtZ'myds, 8. Total 1 7. From the AV. my notes 
show 53 forms (from 16 stems): as, jdmid's, ii.7.2 ; cf. xvi.5.3-5 ; 
8.3-5. Some of the forms are ablatives: as, Jcrshyd's, x.5.34. 

Transitions from C : the Rik shows none ; but the AV. has 
viliptid's, xii.4.44. 

0. 1. The only example is dhiyds (7 times). 

2. Here belong: atharias, iv.6.8; nadias, 2; nishtigrias; 
prpanias, i.71.5; meshias,2; smias, x.101. 3 ; dpias (masc. stem, 
dpia), vi.67.9. In i.121.3, ndkshad dhdvam aruni'h ptirvidm rd't, 
we seem to have a contracted form, aruni's, for -las. 

Genitive Singular Neuter. 
A. The only example is bhti'res (16 times). 

C. There is one example: (vdyasas) suddhias, viii.48.1. 

The Locative Singular. 
The principal terminations of the L.s.m.f.n. of series A are -d 
and -au: -d occurs 126 times; -ait, 272. 

1. The ending -d is used : 

1. before consonants (109 times) ; 

2. before an initial vowel with which it is fused (3 times). 

3. It is never used before vowels (except u-, H-) with hiatus. 
II. The ending -au is used : 

1. before vowels without hiatus, in the form -dv (50 times) ; 

2. at the end of a pdda (150 times). 

The existence of these forms in -d has often been noticed ; but 
it is rather surprising to find that they are almost half as frequent 
as those in -au. The first general distinction in their use is like 
that in the dual (p. 340), and is illustrated by ix.62.15, vir ydnd 
vasatd'v iva. 

I. Of the 126 forms in d: 

1. 109 stand before consonants, in the interior of a pdda. 18 
stand before v-, and 30 before other labials; 12 before y-, and 49 
before other consonants. 

2. 3 coalesce with a following vowel, in the interior of a pdda : 
viii.80.1, sdmam dpi srutd'vidat; vi.12.25y x.101.115. 

3. In 9 cases the sarhhitd, following doubtless the oral tradition, 



386 C. R. Lawman, [i and *-stems. 

has -d u- ($-). The padakdra treats these as he does apparently 
similar cases (Prat, ii.9), and writes -au u- ($-). The references 
are: i.27.66; 36.17d; 124. lab: iv.(>.4ab; 3d.3b: v.l.Qb: vi.26.lfd': 
xAl.ld; lll.Zcd. 

4. "We find -d five times exceptionally at the end of a pdda, but; — be it observed — 
only in the case of two words, devatdtd (vi.4.Ia, before y-: vii.2.56) and sarvatdtd 
(i.94.156: iii.54.196: coalescing graphically with i'-, v.69.3c). Bevdtdt-d, as instr., 
makes perfectly good sense in the first two passages ; in the last three, a dative is 
so appropriate, that Gr. has already on this account suggested the reading 
sarvatdt-e, Wb. 1490. In i.65.4c, yond (g-) is, to be sure, at the end of a pdda; but 
it is aksharapankti: so kukshd' (»-), ix.109.186. 

II. Of the 272 forms in -au: 

1. 50 appear as -dv (35 in the interior of a pdda, and 15 at 
the end of a or c). An a- follows in 30 instances; an i-, 
i.104.7; 176.5: vi.27.1,2; 47.20: vii.18.3; 48.2; an iva, in ii.16.1 : 
viii.92.3: ix.62.15: x.64.6; 97.6; an?-, in vi.15.2 ; an r-, in iii.55.3 ; 
62.18: vii.38.2: ix.13.9; 39.6: Val. 4.2; an e-, in ii.28.7. 

2. 156 are at the end of a pdda (mostly pddas b and d; or a 
and c before consonants). If we add the 15 at the end of a or c 
before vowels, we have a total of 1 71 at the end of a pdda. E. g., 
see iv.16.9 ; the occurrences of djati,, 14 out of 18 ; gabhastau, 6 ; 
abh'ishtau, 10; gdvishtau, 6 out of 7; vd'jasdtau, 12 out of 16, 
v'mshtau, 12 out of 16 ; pti'rasdtau, 7 ; sdta4, 12 out of 17. 

3. There remain 66 cases in which -au is used before a consonant in the 
interior of a pdda (before »-, 6 times ; other labials, 14; y-,1; other consonants, 
39). This happens oftenest with certain frequent words : prasitau, 4 ; yonau, 10 : 
aumatau, 12. It is safe to say in general that many of these occurrences belong 
to younger parts of the Rik-text: thus, 21 (nearly one-third) are in the tenth 
mandala. The passages are as follows : i.31.15c; 48.6rf; 70.4a (Aufrecht calls it 
" Plunder," in the German sense) ; 98.1a; \4A.2b,1d; 162.96 (the horse-sacrifice ; 
late — see Gr., TJeb. ii.452): ii. 13.96 (unclear and mystical); 30.5c: iii.1. 76,21c; 
59M,5d: iv.16.10c; 25. Id; v.30.13d (12-15 are a later addition); 32.7c; 34.6a; 
42.16a"; 43.15a' (these two are refrains); 87.6c (evayd'marut hymn): vLl.lOd; 
16.41c; 24.9d; 26.16; 40.3a; 46.8a; 47.13a (the contents of these two and their 
place in the mandala betoken later date) ; 64.5a'; 73.26: vii. 1.20c (19-20 are a 
separate fragment); 20.8c; 36.8a:; 41.4a'; 46.46; 60.116; 50.3a; 104.13d (these 
two are real Atharvan hymns): viii.3.2a; 48.12a'; 52.126 (later addition); VaM. 3.1, 
ni'pdtithau: ix.72.76; 84.1a"; 97.19d: x.8.3d; 10.76; 14.6c; 25.9d; 29.26; 34.14d 
(gambler's song, end); 38.1a; 44.9d (a late verse, according to its contents and 
bad grammar); 46.66; 64.11a; 85.24c (Sury&'s wedding) ; 88.7c; 101.36; 102.1c 
(does not belong to the hymn); 131.7a (late); 160.4c,5c; 123.6d; 165.46; 
178.2d, bis; the contents of the last three hymns (Gandharva, Peath-dove, 
Tarkshya) show that they are late. 

The termination -au of the L.s. of J-stems has been a crux to the 
comparative grammarians. They say that it is a simple transition 
to the M-decleusion — a formation after the analogy of the w-stems ; 
but since the locatives s. of 2-stems are eight times as frequent as 
those from ?<-stems, this theory will hardly pass muster. A 
complete examination of the forms of the L. in the Rik is so 
suggestive that I cannot forbear saying something about the 
historical relations of the different forms to each other, and the 
genesis of the same. But since this lies beyond the scope of this 



L.s.] Noun- Inflection in the Veda. 387 

article, I would expressly disclaim any tone of categorical assertion 
in the matter. 

The L.s. of the word sd'nu stands in the text in five different forms. The stem 
is " strengthened " by n and the ending is added : as, sd'nu-n-i, i. 155.1 ; or, it is 
strengthened by vowel-increment and the ending is added : as, varshishthe adhi 
sd'nav-i, ix.31.5 (17 forms from 8 stems support this); [or, from this form the 
ending is simply dropped before a vowel, to avoid a cacophonous combination fcf. 
p. 357 and Prat, ii.35) : as, vr'shd pavitre adhi sd'nav-y avyaye, text sd'no, ix.86.3;] 
or, the ending is dropped and the preceding diphthong receives in compensation a 
second increment (cf. Kuhn, in his Zeitsch. xviii.361) : as, sd'ndv adhi, x.123.2,3; 
with this, finally, the usual form is identical : as, adhi sd'nau ni jighnate, i.80.6. 
For the dropping of the ending there are abundant analogies. There are 133 
L.s.n. forms in -man, to 67 in -man-i. Of. camu-i, camu'; tanu-i, tanA'. 

j The facts from the Veda seem to point to a closer parallelism of development 
between the i- and the M-stems in regard to the forms of the L.s. than has yet been 
claimed for them. The analogy of the M-stems, as well as of the dative-locative in 
Latin (ovei, from *avay-i, avey-i) and Greek (irdkii, from iroley-t), makes it probable 
that the original form of the L. of «-stems was made in the same way, with guna 
and ending: thus, agnayi. "With the dropping of the ending took place 
compensatory lengthening: as, agnd'y. Here the lines of development diverge 
slightly. In s&'nav-i the palatal vowel of the case-ending, i is clearly distinguished 
from the preceding labial v. In agndy-i, on the other hand, case-ending and 
thematic final are homogeneous, and (as the word becomes to the apprehension of 
the speaker agnd-yi) both are lost together, so that we have agnd'. 

Now as touching the relation of agnd' to agnail — -there are three possibilities. 
Bach may be a development independent of the other ; this in view of the facts 
few will wish to maintain. Secondly, agnd' may come from agnail ; so Benfey, 
Gram. p. 41 end, and Kuhn in Zeitsch. xviii.366. Thirdly, agnail may come from 
agnd'. When we consider the -an of the dual and perfect (devau, paprau), which, 
beyond all peradventure, is a mere phonetic outcome of -a, it is hard to assume 
here a phonetic change exactly the reverse, without reason. But not only the 
analogy of the dual and perfect aided in this change ; the frequently following 
labial initials would impart their own coloring to the preceding -a (cf. Rik vi.46.1, 
sdta" vd'jasya, with the SV. and "VS. variants, sdtau v-, p. 388); and the analogy 
of the locatives of M-stems (though rare in the Rik — about 50), when standing 
beside those from «-stems, as in i. 162. 9, svdrau svadhitau, would help to fix the 
form in -au. Although the results were the same, the processes were quite 
different. I would therefore set up the series thus : 

agndy-i sd'nav-i 

agnd'(y) sd'ndv 

agnail sd'nau 

Our last form and last but one (agnail and agnd') are abundantly exemplified. 
Perhaps there is some evidence of forms like agnd'y and agnayi. Why may we 
not recognize an instance of elision and crasis in yoneva, for yondy iva, x.101.11, 
as well as in agneva for agner iva, AY. vi.20.la? Again, at the end of i.l04.7d, 
md' no akrte puruMta yondv | indra . . , it is far from unlikely that the oral 
tradition was yondy indra or yond indra. The sandhi is merely graphic here, as 
it is invariably at the end of a pdda. The diaskeuasts almost always wrote the 
sandhi just where it ought not to be (end of a and c) — with perverse uniformity. 
Cf. vi.21Adb.2ab. It is likely enough that the syncopated pdda, vi.44.9rf, once 
sounded thus : dhdnasya sdtd'y o asmd'A avidiihi. Before the pause, the w-sound 
was lost, as regularly in the later language in hiatus (Benfey, Oram. p. 53, Bern. 
1). Now it is confessedly a piece of arbitrariness on the part of the padakdra to 
write -au in the dual before u- (p. 341) ; for the oral tradition unquestionably had -d. 
Was it not equally arbitrary on the part of the diaskeuasts that they filled out 
the hiatus with a v (yondv, sdtd'v) rather than aw? And se, one may ask, is it 
due to anything more than the following of mechanical rules (such as are properly 
applied in yd urd'v antarikshad', v.52.7) that Hrrnau has been put in the pada- 
text of i.27.6, rather that urmai 1 for it may once have been spoken sindhor urmd'y 
updkdy a'. Cf. Delbriick, Verbum % 204. These are mere suggestions. 



888 0. ft. Lanman, [i and z-stems. 

Forms like agnayi seem less doubtful. Gr. proposed ekeing out djau to a 
dactyl, dj&vi, in x.75.96 and i.112.10. In x.75.9, b may be catalectic — and a, c, and 
d, full jagatis ; so in x.63.14a (vd'jasdtayil). In i.112.10, a dactyl is certainly 
needed ; but I would read it thus : sahdsramidha djayi ajinvatam. In like manner, 
x.46.6, pdrivtto yonayi sidad aniah, text -au. In vi.33.4c, syncopation is not in 
place (as it is in vi.26.la, owing to the pause in the sense). I am therefore 
inclined to pronounce suarshdtayi ydd dhvaydmasi tvd (text -id). In x.150.4, we 
have agnim mdho dhanas&ldv aham huve | mrdikam dhanasdtaye. In every other 
verse of the hymn, pdda d repeats part of c (as so often happens in pdda c of 
atyashti, i.134, etc.). We should, then, expect a locative in d; and besides, it 
would be unnatural to find the same relation ('to favor,' 'to the getting of good') 
differently expressed — now by the ace, and now by the dat. — in two consecutive 
words. Does not -taye stand in place of an original dhdnasdtayit Cf. viii.3.5a 
and i.4.9c. 

Locative Singular Masculine. 

A. 1. Here belong possibly djdyi, i.112.10; ydnayi, x.46.6. 
See above. 

2. There are 40 m. forms (from 7 stems) : agnd', 11 ; djd', 6 ; 
■Armd', 5; kulcshd' , 2 ; ghr'nd, vi.3.7; ydnd, 13, and ix.86.6 (Saman 
variant, ii.237, -au), and once in AV. (a Rik-passage), ix.10.10; 
surabhd', v.1.6. Cf. gara, Y<j. x.4, ' gira'u.'' 

3. There are 101 forms in -au (from 2V stems) : agnaA, 20 ; 
ddrau, 3; aratau; aratnaH,; djau, 18; dnau; firman, 3; 
gdbhastau, 6 ; girau; ydpatau; jtrau, or f. ; trkshau; dhvasdnau; 
ndmucau; ni'pdtithau; panau; prtsutati; medhidtithau; yojnd- 
patau; ydnau, 29 ; rdthavitau; vdnaspdtau, 2 ; palmalati,; 
sdnitau; sd'rhvaranau; syd'marapmau; svddhitau. From the 
Atharvan I have 42 forms (from 18 stems). In v.31.5, we see the 
later style of sandhi, pHrudgnd'v utd, for which the Rik would 
have -nd' utd. In iii.18.3, the unoompounded pdtau is merely a 
false variant of the Rik jdne, x.145.4. 

B. C. There are no examples. 

Locative Singular Feminine. 

A. 1. Here belong possibly su.arshdtayi, vi.33.4, and dhdna- 
sdtayi, x.150.4. See above. 

2. There are 85 f. forms in -d (from 22 stems) : avdnd; djd', 
i. 116. 15; uditd, 10 ; kshetrasdtd; goshdtd,2; tokdsdtd, vi.18.6 ; 
devdtdtd, 10; dyumndsdtd; nd'bhd, 20; nr'shdtd; nernddhitd, 4; 
pltd'; prdbhrtd; niedhdsdtd, 3 ; yd'mahutd, x. 1 1 7.3 ; vhcshtd, 
x.41.1; pu'rasdtd, 5; sdmrtd, 2; sarvdtdtd, 6; sdld', 5 and 
vi.46. 1 (SV. i.234 and VS. xxvii.37, -au) ; srutd'; suarshdtd, 6. 
AV. has nd'bhd, vii.62.l ; and the Saman variant of Rik viii.18.7 
has pdmtdtd, i.102. In Yc. xxxi.18, we find dusitd, ' dtishtau.'' 

3. There are 171 forms in -au (from 49 stems) : dcha-uktau, 2 ; 
djushtau; ddhitau; abhishtau, 10 ; arkd-sdtau, 3 ; drnasdtau, 3 ; 
d'gatau; ishtau,,' 6 ; Upastutau; upetau; urukshltau,; etau; 
'eshtau; gdvishtau, 7 ; tokdsdtau, x.25.9; durmatau, 2 ; devdvitau, 
6 ; devdhHtau, 2 ; dyumndh/Atau, 2 ; dhdnasdtau; ndvishtau; 
nd'bhau, 2 ; nrtau; nr'pitau; pdrishtau; pdretau; pita-u,, 2 ; 
pushtau, 2; purvdhutau, 5; prdnttau, 3; prdbhUtau; prdbhr- 
tau, 2 ; prdsitau, 4 ; bhdrahMau; medhdsdtau, 3 ; rdtau, 5 ; 



L.s.£] Noun- Inflection in the Veda. 389 

vasatau, 2; vd'jasdtau, 16; viushtau, 16; ffuptau; pA'rasdtau, 
7 ; prushta'u; sdmdrshtau. 3 ; sdmitau; sdmrtau, or m., 3 ; sdtau, 
17; sumatau, 14; sufishtau; suastau. My notes show 19 true 
AV. forms (from 7 stems). For -uditd of the Rik, vii.41.4, the 
AV. (but not the VS., xxxiv.37) has the later form -dditau, iii.16.4. 
4. I am not sure but that the following are transition-forms 
from A to B, especially as vedi becomes vedi in the post-Vedic 
literature. In this case they should be added on p. 371. They 
are: viM, vi.1.10; v'edy asyd'm, p. vedi iti, ii.3.4 ; cf. Prat, i.28: 
ii.35. In iii.51.7, prdniti might be a locative. 

B. 1. Here belong 13 forms (from 7 stems): (a) devid'm; 
prthivid'm, 5 [and from the AV., nardcid'm, v.31.4; striyd'm, 
vi.11.2; 70.1-3]: — (b) dsiknidm, 2 (and AV xii.2.20) ; ja/ind'- 
vidtn; pdrushnidm/ uchdntidm, 2 ; ydcidm [and varand'vatidm, 
AV. iv.7.1]. 

2. Here belong 34 forms (from 9 stems) : (a) aranydnydlm,; 
dshtryd'm / catvdrinpyd'm • prthivyd'm, 26 [AV, dsandyd'm, 
xiv.2.65 ; gdyatryd'm, xiii.3.20 ; jyaishthaghnyd'm, vi.110.2 ; 
devyd'm, vi.136.1 ; prthivyd'm, 40] : — (b) drshddvatydm ; 
yavyd'vatydm; rd'trydm; vasd'vydm (stem -vi; cf. jahnd'vi-m) ; 
sdrasvatydm [and from AV., 14 forms, from 5 stems]. 

Transitions from A: 1. p&ramdhidm ; bhU'midm, 2 [and from 
the AV., prshtid'm, vi.102.2; bhti'tidm, xii.1.63 ; 4.44,46; bM!mi- 
dm, 15 times: as, ii.9.4; ydnidm, v.25.8; vedidm, xii.3.23]: — 
2. bhrtyd'm; yuvatyd'm; sdmgatydm [and from the AV, 
d'kHtydm, cittydm, and devdhtitydm, each 17 times, v. 24. 1-1 7; dv- 
ydrn, 3; bhU'mydm, 4; v'edy dm, xi.1.24: xix.33.3 ; sdmitydm, 2]. 

Transitions from C: diUid'm, see p. 373 ; priyd'm, AV. xii.1.63. 

3. The ending is dropped : gauri', ix.12.3 ; sarasf, vii. 103.2. See 
Prat, i.28: sdptamikau ca pHrvau (i. e. ikdrokdrau, pragrhyau). 
Cf. vedi, prdniti, A 4. The commentator to Ath. Prat, i.74, q.v., 
cites dshtrf, vi.27.3 (Rik x.165.3, dshtryd'm); vrvi', xviii.1.32; 
maihi', xviii.1.39 (a queer variant of Rik x.31.9). To these the 
editor adds dbhihruti, vi.3.3. 

C. There is no example. 

Locative Singular Neuter. 

A. The only example is apratd ', viii.32.16. The AV. has sap- 
tdrafrn.au, ix.5.15; Brh. ar. up. iv.2.3, akshini. 

B. C. There is no example. 

Vocative Singular Masculine. 

A. The stem-final always shows an increment. Schleicher's 
Vedic vocatives without increment are transition-forms to B. 
Thus, the voc. darvi is from stem darvi', Rik ddrvi. The form 
occurs 1105 times (from 72 stems). Examples are: agne, 524; 
dgne, 275; sum, 799; dghrne, 13; kave, 26; pate, 35; with 
brhas-, 32; brahmanas-, 17; vanas-, 14; sat-, 13; viearshane, 
11; sakhe, 11. In ix.80.1, we have br'haspdte. With ma'he- 
mate cf. mahe-nadi. 

VOL. X. 53 



390 C. R. Lanman, [i and t-stems. 

Transition from B : sobhare, 3. 

Transition from the M-declension is seen in the vocative s. of 
the compounds of poets : pdvakag,oee s 2 ; bhadrapoce, 4 ; guhra- 
poce; cf. sahasraeaksho, w-stems, V.s.m. This is doubtless done 
by the poet in order to avoid an unfamiliar, or otherwise never 
occurring form. Cf. p. 377 end. The prescribed form is sujyotis. 

B. There is no example. 

C. In i.141.8, we have dakshi, p. dhakshi (Prat, iv.41). Other- 
wise BR. 

Vocative Singular Feminine. 

A. Here belong 27 forms (from 11 stems). They are : adite, 
15 ; anumate; asunite, 2 ; ishte, 2 ; upamdte; rjite; bhilme; mate; 
yuvate ; satyatdte ; sua figure. The AV. has darve, iii.10.7 
(cf. x.4.13) ; VS., darvi. 

Transition from B : oshadhe, 2. 

B. There are 117 forms (from 38 stems). They are: arunydni; 
arushi; arjuni, 2 ; ayvdjani; apvdvati; indrdni; urilci; rtdvari; 
gornati; devi, 23; navyasi; ndri; parushni, 2; prthivi, 11: see 
V.d.f. ; pravatvati; bhdrati; maghoni, 9 ; mahi, 6 ; mahini; 
mdnushi; rdtri; revati, 2 ; vananvati; vdjini, 2 ; vdjinivati, 8 ; 
vicdrini; vibhdvari, 8 ; vrshdTcapdyi; pitikdvati; $utudri; 
fdrapatni; sarasvati, 16 ; sahasvati; sinivdli, 2 ; sunari; s-dnr- 
tdvati; sunrtdvari; hid 'dikdvati. 

C. 1. No example. 2. ardyi; mahenadi; yarni. AV. vii.115.1, 
lakshmi. 

Vocative Singular Neuter. 

A. B. C. There is no example. 

Nominative, Accusative, and Vocative Dual Masculine. 

A. The analysis of these forms is doubtful. Schleicher sets 4 
and -H = -yd and -vd; but there are no traces of this origin. For 
lack of something better I have put 4 = -i-i, p. 366. Friedrich 
Milller, Sitzungsberichte d. kais. Ak. d. Wiss. (Vienna, 1860) 
xxxv.55, compares the duals devd, kavi, bhdwCh with the singulars 
devd, kavi, bhdnib, and recognizes therein a reduplication of the 
end of the word, symbolical of its dual meaning. The same 
explanation applies unfortunately to the plurals neuter bhuvand, 
ffuei and tri, and puru, as compared with the singulars in -&, 4, -it. 

There are 340 forms (from 72 stems). Disregarding accent, 
examples are: indrdgni, 78 ; pati, 33 ; pubhaspati, 21 ; hari, 78. 
For pdpva4shti, see p. 380. 

The final is short in sakshdni, x.32.1 (cf. viii. 22.15). So we 
have asurd, etc., patird, vird, p. 342 ; and jigatnib', vii.65.1 ; and 
others below, p. 391, B, and 392 top. 

B. The formation of the stem of mddhvi — it occurs 7 times, as 
epithet of the Aevins — is uncertain ; but md'dhvibhydm, VS. 
vii. 11, leads us to class it with Ndrni, rd'shtri, etc. (p. 367 end), 
as a B-form. 

C. 1. Root-words: manani a; sadhanid. 2. rathid, 3; rathid, 8. 



KA.V.d.f.] Noun- Inflection in the Veda. 391 

Nominative, Accusative, and Vocative Dual Feminine. 

A. The form occurs 23 times (from 15 stems). They are : 
itdsiXti; Uti'j r'jiti; jdmi', 2 ; ddrvi; dyd'vdbhil'mi, 2 ; voc, 3 ; 
dhdraydtkavi; purarhdhi; bhuji 1 '; yuvati', 3 ; vdsudhiti, 2 ; puci; 
sdyoni; suprdturti; srutf. 

B. Series B and C are here distinctly separated from each other 
in the Rik. The Sanskrit form -yau, or even its Vedic equivalent, 
-id'' or -id, is utterly unknown to the Rik (p. 370). It is really no 
B-form, but a transition to C, which at last became universal, and 
whose beginnings we see in the Atharvan (cf. C 2). 

There are 368 forms (from 76 stems). Examples are (vocatives 
included): urvi ', 20; devi', 13; dyd'vdprthivi', 65; voc, 15; 
prthivi', 20; prthvi', 5; brhatt, 10; mahi', 27; yahvil , 6 and 
x.93.1; rddasi, 87; samici', 11. In i. 162.21, hart te yAnjd pr 'shati 
abkutdm, we cannot interpret pr' shati — with Gr., Ueb. — as 'the 
two spotted gazelles ' of the Maruts, for their team consists of 
many. We must assume non-agreement in form between hdri 
(masc. — vr'shand, iii.35.5 etc.) and pr' shati (fem.), and take them 
together : ' The two spotted bays (of Indra) are become thy 
companions.' But pr'shat-i may be N.d.n. See stems in ant, and 
cf. sujdnman-i dhishdne, i. 160.1. The form aksM ', furthermore, 
occurs six times. It is perhaps best referred to B on account of 
ahshi'bhydm (x.163.1), akshids (AV. vi.127.3), and -yds (24.2). 
Line 13, p. 371, should be corrected accordingly. 

Transitions from C : kshoni', viii.7.22 ; 88.6: Val. 4.10; nadt', 
i. 135.9; for gaurf , cf. p. 373, 389. 

The final appears to be metrically shortened in the seventh place 
of ii.31.5c (stushe ydd vdm prthivi), and of ni.54Ad. In x.93.1, 
mdhi dydvdprthivi, and in iv.56.5, prd vdm mdhi dydvi abhi, 
perhaps the double consonant made a written -I needless. In view 
of the numerous analogies (p. 390, A), it is better to take them 
as real duals in -i, rather than to assume a false reading (p. 377) 
or the like. 

C. 1. Root-words: abhipriyd, 2; ghrtapriyd; AV. viii.2.14, 
abhipriyau. 

2. Here belong : nad'td; naptid; yamid, 2 ; sakthid, 2 ; srnid, 
x.106.6; from oakrf, cakriyd, i. 185.1 : ii.34.9: v.30.8: x.89.4(the 
Saman variant, cakriyau, i.339, shows that this is dual and not I.s. ; 
cf. Wb. 429, 1761). The Kathaka has cakriyau, xxix.7; the AV., 
sakthiau, vi.9.1 ; and nddyait, vi.l 38.4. 

Transitions from B. These are the rule in Sanskrit. The Rik 
has not a single instance of it. The Atharvan has the following : 
dndiau, vi.138.2 ; phdlgunyau, xix.7.3 ; and, finally, akshyaii, 
i.27.1: iv.3.3: v.23.3; 29.4: vi.9.16, lc(ed. -au!): vii.36.1: xix.50.1. 
The AV. has no example of -id. 

Nominative and Accusative Dual Neuter. 

A. 1. (a) Here belongs p&ci (cakre), x.85.12; possibly, p&ct, 
iv.56.5 ; and from the AV., tigmdheti (pr'nge), viii.3.25. (b) In 



'692 0. R. Lanman, \i and 2-stems. 

RV. ix.97.54, mdhl (vddhatre) appears with shortened final ; and 
so, perhaps, in iv.56.5. 

2. BR. and Gr. put hdrini (pr'?ige), ix.70.7, under hdrita. I 
would rather refer it to hdri. The AV. has dkshini, x.9.14 and 
xi.3.2. 

Instrumental, Dative, and Ablative Dual Masculine. 

A. Instrumental : inclrdgmbhydm, x.128.9 ; hdribhydm, 22 
times, and iv.15.7, and AV. ii.5.1 ; ddmpatibhydm, AV. xii.3.27. 
Datives : indrdgmbhydrn, 4, and v.86.6«, evd indrdagnibhydm ; 
hdribhydm, 2. Ablatives: kukshibhydm, AV. ii.33.4, and pdnib- 
hydm, verse 6. In no case is -bhidm needed. 

Instrumental, Dative, and Ablative Dual Feminine. 

A. Instrumental : vartaiubhydm. Ablatives : pd'rshnibhydm 
and prdnibhydm, x.163.4 ; prdnibhydm, AV. ix.8.21. 

B. Rddasibhydm, D.d.l'., i.136.6; aksht'bhydm, Ab.d.f., x.163.1 ; 
Ld.f., AV. xi.3.34 bis. 

C. In ii.16.3, kshoni'bhydm is I.d.f. 

Instrumental, Dative, and Ablative Dual Neuter. 
A.B.C. In AV. x.10.21, sdkthibhydm is Ld.n. 

Genitive and Locative Dual Masculine. 

A. Genitives : indrdgnids, 2 ; hdrios, 2. Locatives : kiikshios, 
2; gdbhastios, or f., 21 ; hdrios, 2. In ix.58.3, purushdntios does 
the duty of an ablative ; cf. p. 344. 

C. In x.88.17, yajnanios is genitive. 

Genitive and Locative Dual Feminine. 

A. Genitive: yuvatyds, 2. Locatives: jdmios, v.19.4; gavl- 
nios, AV. i.3.6 : v.25.10-13 ; pd'rshnios, vi.24.2. 

B. Genitives: 1. akshids, AV. v.4.10 (MSS. and ed., akshds) ; 
parinr'tyantios, AV. x.7.43 ; rddasios, RV., 13 times ; and in the 
sense of an ablative (p. 344), vi.24.3; from nini' — m. ninid, 
oxytone — we have ninids, x.5.1 (cf. p. 369): or, it may stand for 
ninidyos, like pastios, etc., p. 344 : — 2. dwaspr'thivyds, 4 ; 
rddasyos, 2. Locatives : 1. ardnios, iii.29.2 ; pdtantios, x.24.5 ; 
rddasios, 3; akshids, AV. vi. 127.3: — 2. samtcyds, x.24.5; 
ardnyos, vii.1.1 ; drjunyos, x.85.13 ; akshyds, AV. vi.24.2. 

C. Genitives: onlos, ix.65.11 (cf. Ath. Pr. iii.61) ; cakrios, 2. 
Locatives: on'ios, ix. 16.1 ; 101.14; naptios. 

Genitive and Locative Dual Neuter. 

A.B.C. There is no example. 

Nominative and Vocative Plural Masculine. 

A. The exceptional ary-ds (16) is the sole example of its kind. 
The stem-final is regularly gunated and the ending added. This 
form occurs 523 times (from 109 stems). Examples are : agndyas, 



N.V.p.m.] Noun- Inflection in the Veda. 393 

27; ddrayas, 26; arcdyas, 14; tirtndyas, 15; rf shay as, 26; 
kavdyas, 45; pdtayas, 16; rapmdyas, 16; vdhnayas, 15; vdyas, 
38; pUcayas, 18; silrdyas, 36; hdrayas, 31. In vi.66.11, girdyas 
is possibly an Ab.s.m. ; of. p. 383, 384 end. Gr. proposes giribhyas. 

Transition from B : Sdbharayas. 

Transition from C : dhrayas (p. 372). 

Transition to the w-declension (?). In ii.34.2<*, 'rings' (khddi-n-as 
■=zkhdddyas) would be more appropriate than 'wearing rings' 
(khddln-as) ; but this equation is hardly to be allowed. 

B. Sole example : siri's. 

Transition from C: takvavi's, i.151.5 (p. 373). 

C. 1. Root-words: (a) durddhias, 2; d-ddhias; devdvias, 2; 
padavias; sadhamas; sudh'tas, 4 ; suddh'ias, 15; grdmanlas, AV . 
iii.5.7. (b) agniprlyas; adhvarapriyas, 3 ; abhiprlyas; for 
dirghd'dhiyas and nd'nddhiyas, cf. p. 369 ; paripriyas; suyrtyas, 
2 ; kadhapriyas, 2 ; ganapriyas, i.64.9. 

2. Here belong : ah'ias; dpathi'as (cf. d'pathayas) ; rathias, 3 ; 
rathias, 13 ; for suhastias, ix.46.4, see pp. 368-9. 

Nominative and Vocative Plural Feminine. 

A. We have aryds, f., i.71.3; 122.14: x.'27.Sbis; and of the 
usual form, 290 occurrences (from 66 stems). Examples are : 
drdtayas, 18; dtdyas, 42; krshtdyas, 17 ; kshitdyas, 13 ; jdnayas, 
12; dhitdyas, 21 ; matdyas, 20; rdtdyas, 18 ; vrshtdyas, 12. 

Transitions from B: Sshadhayas, 3; see others on p. 372. 
Transitions from C : ambdyas; arundyas, x.95.6 ; kshondyas. 

B. 1. The uncontracted form is perhaps seen in vd'nias, Val. 
11.35, text, -1?/ but the verse may consist of 12, 11, 12, 12. 

2. The regular form of the later language (devyds) is almost 
entirely unknown in the Rik. Instead of that, we have devl's. 
Forms of this kind occur 423 times (from 166 stems). Examples 
are: dirts (vipas), 1.77.3; 96.3 (Gr., A.p.f.) : cf. d'rids; rtd'varis, 
3; dshadhis, 27 ; devaydntis, 6; devi's, 43 ; pdtnis, 11; puroi's, 
36; bhdti's, 8; maht's, 11; yahvi's, 7; vdsvis, 6; vd'nis, 12; 
sincati's and sincdntis; for dcipvis, i. 120.8, see p. 372 med. 

Transitions from A : avdnis; djd'nis; -dti's; ndktis; nirrtis; 
nlshkrtis ; ptirarhdhis, i.123.6: v.41.6 ; bhU'mis, 2; vipvd- 
krshtis; for prhiis, v.59.7, Gr. proposed prdy inis — better prenay as; 
sdyonis; from the Atharvan : angMis, x.2.1; rshtfs, iv.37.8,9; 
dhamdnis, vii.35.2 ; nd'bhis, xix.53.2 ; pd'rshms, viii.6.15 ; prshtfs, 
x.9.20: xi.8.14; bhd'mis (read -mayasf), xi.7.14. See p. 370 top. 

Transitions from C. There appear to be none in the Rik : for 
rathi's, viii.84.1, see p. 373 med.; for kshoni's, i.54.1, see p. 372 
end. In the Atharvan we have nadi's (but also nadias), iii.24.3, 
and lakshmi's (but also -rntas), vii. 115.4. 

C. 1. Root-words: (a) ddhlas, i.105.7; 105.8=x.33.3; pramas; 
from AV., viddhias, vii. 114.2 ; in the Rik it would be written 
viddhiyas(]).369):—(b)abhipriyas,3; dhiyas,l9; vd'tapramiyas 
(p. 369); priyas, 6; dpriyas, AV. xi.7.19. 



394 0. R. Lawman, [i and ^-sterns. 

2. Here belong: (a) arunias, i. 140.13; enias, 2; kalydnias; 
gaurlas ; nad'ias, 15 ; maydrias ; rathias, vii.21.3 ; vakshias; 
starias; from the AV., ardyias, i.28.4 : ii. 14.3 ; tikshnaprngias, 
viii.7.9; nadias, vi.24.3 : xix.1.1 ; nddlas, x.7.15; mayurias, vii. 
56.7; ydtudhdnlas, i.28.4 : ii.14.3: xix.47.7; lakshmias, vii. 115.3 ; 
vikepias, i.28.4: xi.2.11; 9.14; viastakepias, viii.1.19; sahasra- 
parnias, viii.7.13 : — (b) samudrlyas, Rik x.65.13. 

3. The stem-final is consonantized in the Rik only in starydm, 
vii.68.8<#, and nadyd, vii.50.4. See p. 379 top, and p. 373 med. 
The Atharvan has: apvataryds, viii.8.22 ; nadyds, iii.13.1: xi. 
6.10: xiv.2.7 ; naptyds, vii.82.6 ; nddyds, x.7.16 ; pippalyds, 
vi.109.2; vrkshasarpyds, ix.2.22. 

Transitions from A : yayiyas, x.78.7 ; pubhr'iyas, AV. xx.48.2 ; 
karkaryds (ed. -yds), iv.37.4. 

Transition from B. This has become universal in the later 
language. There is only one certain instance of it in the Rik 
(stri-y-as, vii.55.8). Even in the Atharvan we see only the begin- 
nings of the movement, and the N.p.f. of series B ends almost 
always in -is; thus, devils occurs 19 times as N., and 5 times as A. 
We have as N.p.f. urvias, xi.7.18, and elsewhere urvi's (9 times). 
The longer form is preferred on account of the metre ; and perhaps 
this gave the first impulse to the movement. Here belong : (from 
RV.) prganias, x.61.8, and suparnias, 88.19: — (from AV.) striyas, 
iv.5.2 : vi.138.5; urvias, xi.7.18; rudatyds, 9.14. Cf. p. 372. 

Nominative and Accusative Plural Neuter. 

A. There is no example of a vocative. The N. and A. have a 
series of forms entirely parallel with those of the stems in a, in u, 
and in man. 

1. Here belong : kr&dhmi (mandnsi), vii.56.8 ; pact (havyd!), 
56.12; and the numeral tr't! , 21 times. In vi. 44. 14 and vii.23.3, 
(vrtrd'ni) aprati', the pada has -i. 

2. Both texts have -l : (vrtrd'ni) aprati, i.53.6: ix.23.7; asthUrl 
(gd'rhapatydni), vi.15.19; jdmi (brdhmdni), vii.72.3 ; (d'yudhdni) 
x.8.7 ; bhtl'ri (dnnd, tdnaydni, nd'ma, etc.), 15 times, and 
x. 120.5,6 ; pdmi (ptirvid'ni, BR.), iii.55.3 ; surabfu (mtikhd), 
iv.39.6 ; and from the AV., aprati, vii.50.l5i* • 93.1; 110.1; 
bhU'ri, v.2.6 (Rik, v. 1.); 11.7: vi.1.3; mdhi (vdrcdnsi), iv.22.3. 

3. The Rik has: aprati'ni, 5; bhlX'rtni, 5 times, and i.l65.7c 
(bhU'ri in a) ; pucini; surabhi'ni, {havyd', -yd'ni), 2. The AV. 
has : dkshini, iv.o.5 (Rik, akshd'ni) ; dsthini, ix.5.23 : xi.8.12 : 
xii.5.70. In Rik i. 149.4, the metre demands the longer gram- 
matical form, tri'ni — ^text, tri'. 

B.C. There is no example. Paradigm : jalapi'ni. 

Accusative Plural Masculine. 

A. We find ary-ds 7 times (in 602.7, it is N.s.m.). The usual 
pada-iorm -in occurs 84 times (from 31 stems). It appears in the 
samhitd as -inp in i.173.8, and as -inr, 41 times. That is, in one- 



A. p.m.] Noun- Inflection in the Veda. S95 

half of all the instances, we have the direct representative of the 
organic form -ins (p. 346). 1. The form -Mr appears only in the 
interior of a pdda. It occurs (a) before a vowel: viz., before a-, 
26 times: as, ix.l07.19<?/ before iva, in i.52.5; 141.11: iii.45.3: 
iv.30.15: viii.35.21 ; before i-, i.109.3: iii.38.1; before u-, viii. 
26.10: x.69.12; before o-, v. 41.8 ; 42.16; furthermore, (b) before 
y-, iv.35.7; before v-, i.127.7 and vi.39.2; before h-, i.184.2. For 
the last four cases the Prat., iv.29, makes special provision. 2. In 
28 of the other 42 cases, the form appears as -in at the end of a 
pdda: as, giri'n \ {id, x.138.2 ; and in the interior of a pdda, as 
-in, 10 times — before p- (5 times), k-, t-, n-, y-, v-; a,s -in or -int 
before s-, i.144.3: iii.4.1 : vii.57.7; and once as -in before j-, 
vii.3.8. 

Accusatives p.m. are found from the stems: agni; djdmi, 2; anji; atithi; 
adabdhadhtli ; apidhi; dsushvi, 2; dji; dpi; ishudhi; tidadhi; Urmi; r'shi; kavi, 
2; girt, 4; jdmi; nidhi, 3; pani, 9; paridhi, 4; pradhi; rapmi, 6; vddhri; 
vanaspdti, 11; vijdmi; vidti, 2; sdkhi, 5; sdlpati; saplarshi; suri.U; sti,2; 
smaddishti. 

In vimticydvdyo avasd'ya dpvdn, i.104.1, the stem v'l is irregu- 
larly gunated before the ending is added. See p. 384, and girdyas, 
p. 383. 

C. 1 . (a) durddhias ; dUdMas, 6 ; sadhamas. (b) supr'vyas, 
viii. 8. 17. 2. ah'ias; dushprdvias; rathias, 2. As a transition- 
form — from A — cf. avdntyas, AV. Paric., quoted by BR. 

Accusative Plural Feminine. 

A. We find ary-ds 4 times. The usual form in -is occurs 122 
times (from 42 stems). In i.140.9, I take vartani'r as N.s.f. — p. 
377. For drdtis, N.s.f., vi.4.5, see p. 376. For supposed forms in 
-ias — text -is — see p. 371 ad fin. 

The stems are : an&nubMM ; dbhimati, 5; abhipasti; abhi'ti; drdti, 16; avani, 
6; avyathi; a$ asti, 5; ishti; titi, 3; rshti; krshtl, 16; krshnayoni; kshiti, 8 
gavyuti; carsJiani, 7; jdmi, 3; tvishi; dnrmati, 2 ; pakli, 3; pankti; pwramdhi, 1 
pushti; prtsuti, 2; prshti; bhu'mi, 3; mail, 4; marici ; yuvati; vankri; vartani, 
iv.19.2 ; vasati; fakti, 2; samiti; sdyoni; sdti; sukshili, 3; sumati, 2; sushtuti, 
2; siuti; srakti; heti. 

The A. p.m., vdyas, would seem to support citrdtayas, x. 140.3, 
as A.p.f. To this we may add piicayas (dnu svd'h), AV. v. 1.3. 
Cf. p. 384. 

B. The form coincides with that of the nom., and with that of 
the ace. of A. It occurs 310 times (from 106 stems). Examples 
are : ddevis, 10 ; dshadhis, 24 ; dd'sis, 9 ; pdtnis, 7 ; pHrvi's, 40 ; 
pr'shatis, 1 ; brhati's, 12; rnahi's, 18. In viii.50.16, d is cata- 
lectic ; we need not assume the uncontracted form ddevius — text, -2s. 

Transitions from C : aruni's, i.112.19: iv.2.16; ydtudhdni's,AV. 
iv.18.7c (the metre would allow the reading -ias). For nadio, 
ix.9.4 — Gr., nadi's — see p. 373. I take kshoni's, i.57.4, as N".s.f. ; so 
also QokatHs, x.146.3 ; and as N.s.m., rathi's, iii.30.11. 

C. 1. (a) ddhias, AV. vi. 131.1. (b) dhiyas, RV. 46 (and AV. ix. 
2.25) ; prlyas, 12; bhiyas, AV. iv.31.7 (Rik bhiyam). 



396 0. R. Lanman, [i and 2-stems. 

2. (a) ashtakarnias ; Mlds'ms; hhdrias; dehias, 2 ; nadias, 9 ; 
naptias; meshias; yamias; ydtudhdmas, 2 ; starias. From the 
A V., ardyias, iv.17.5 ; 18.7,8; nadias, vi.12.3: vii.38.5; naptias, 
ii.14.1 ; ydtudhdmas, i.28.2 : iv.9.9 ; 20.6: xix.36.2 ; 39.1,5,8; 
sambddhatandrias, x.2.9. (b) samudriyas, RV. i.25.7; 55.2. 

Transitions from B: yahvias and suparnias (but see p. 372); 
stri-y-as. 

Instrumental Plural Masculine. 

A. The ending is added directly to the stem : as, agn/trbhis. The 
form occurs 258 times (from 40 stems). 

The stems are : agni,10; djdmi; anji, 10: ddri, 43; arci, 2 ; dvi, 4: avyathi; 
dghdti,ovi.; upabdi, 2; ilrww, 5; rtddMti; r'shi, 10; kavi,7; jdmi,3; dravatpdni; 
nidhi; pani; pafhi, 28; p<m, 2; j)«»»; prdshti; bhu'ri, 3; raj/«; raprol 36; 
vanaspati ; vahni, 5 ; vi, 9 ; vidupavi, 2 ; vtdupani ; fuci, 2 ; saAAi, 1 5 ; sopjfe' ; 
sdpti; sd'dhadishti; suditi, 5; siiri, 14; svdyutcti ; sveddnji; hart, 23; hiranya- 



C. Here belong only three words : suddhi'bhis, 2 ; ganaprVbhis 
(p. 372); and Mranyavdpibhis. 

Instrumental Plural Feminine. 

A. The form is like that of the m. : &s,jdni-bhis. It occurs 365 
times (from 48 stems). 

The stems are : acitti; dcha-ukti, 3 ; dbhishti, 1 ; arishtatdti; avyathi; afdni ; 
asdmi; ishti, 2 ; &ti, 100; r^'ife'; rshti, 6; krshti; gikti; citti, 5; citt»; /dm, 
5; jdroi, 4; di'dhiti; devaviti; devahUti, 4; dyumnaMti; dhiti, 27; ndma-ukti ; 
paribh&ti; praniti ; prdpasli, 8; wcrfj, 37; yuvati, 2; ratfi, 3; vd'jasdti ; vishti: 
vrshti, 4 ; fakti, 2 ; (atabhuji ; (amtdti ; fasti ; freni ; sahUti, 2 ; suntti, 2 ; sumati, 
6; nvrili, 18; sufasti, 6; mshtnti, 2; svdyukti; svdvrkti; suasti, 7, and in the 
refrain of the Vasishtha-hymns, 75 times; sudpi; havydddti, 4. 

Transition from B : ardnibhis, 2. 

In nine instances, the word Utt is shown not only by the sense, 
but also by the forms of the adjectives in concord with it, to be 
Lp.f. : as, tvd'bhir 4ti', ii.20.2 ; and in iv.29.1, it is parallel with 
vd'jebhis. 

B. The form is distinguished from that of A only by the stem- 
final: as, dnvt-bhis. It occurs 113 times (from 32 stems). 

The stems are: dnvi; drusht ; drvali; apmanmdyi; ashtd'padi; dtmanvdti; 
d'yasi, 2; oshadhi, 6; jaratt ; javini; tdvisht, 13; dvnyC ; dhautarl; navyast, 4; 
pdtni ; pArvt', 4 ; pr'shati, 5 ; pravdtvati ; brhati' ; bhd'rati ; mahi', 4 ; yahvi' ; 
rd'tri; vamri' ; var&trt; vdsvt; vd'nl; vd'pi, 3; fad, 36; fatini, 4; (ami, 8; 
sahasrini, 4. 

C. 1. dMbhis, 37 (and AV. v.20.8); pribhls. 2. halydni'bhis; 
kshont'bhis, 2 ; nadi'bhis; napti'bhis. 

Transition from B. The word stri-bMs I would put here rather 
than directly under B, on account of the accent. Herein it fol- 
lows the general rule for monosyllables of series C. 

Instrumental Plural Neuter. 
A.B.C. There is no example. 



D.Ab.p.m.] Noun- Inflection in the Veda. 897 

Dative and Ablative Plural Masculine. 

The general ending, -bhyas, is affixed directly to the stem : as, 
vi-bhyas. The two cases are coincident in form, but are grouped 
here according to the exegesis. 

A. Datives. There are 44 forms (from 9 stems). The stems 
are: r'shi, x.14.15; rshibhyas, Val. 11.6 (the accent is perhaps a 
mere blunder of the tradition) ; ddpdvani; pdti, 3 ; ydti; vi; 
ptici; sdkhi, 21, and iii.31.15 ; sani, 4; stlri, 9. 

Ablatives. There are 17 forms (from 12 stems). The stems are: 
agni; dtri; Jcavdri; girt, 3 (Gr. proposes gir'ibhyas, vi.66.11; see 
N.p.m.) ; pant/ pavi ; pldpi ; vdnaspdti ; vi; sdkhi, 4; sani; 
sthiv't. 

C. The only example in the Rik is the dative, rta-ni' -bhyas, 
ii.27.1 2, with shortened i'. Cf. sendnibhyas and pvanibhyas, p. 372. 

Dative and Ablative Plural Feminine. 

A. Datives : kshitibhyas; earshanibhyas, 6. 
Transition from B: nd'ribhyas, viii.46.8; -bhias, i.43.6. 
Ablatives : ahhatibhyas; kshitibhyas, 2 ; earshanibhyas ; jdni- 

bhyas, x. 183.3 ; aiigi'dibhyas and dhamdnibhyas, AV. ii.33.6. 

B. Datives: (a) tdvishibhyas ; from AV., gandharvdpatni- 
bhyas, ii. 2.-5 ; papvati'bhyas (sic), v.8.8 : vi.75.2,3 : — (b) at the end 
of catalectic pddas (Gr., -bhias) : dshadhibhyas, ix.11.3 ; vrshan- 
ydntibhyas, 19.5; ghoshinibhyas, AV. xi.2.31a/ kepinibhyas, b. 

Ablatives : (a) dshadhibhyas, vii.50.3 : — (b) dshadhibhyas, 
ii. 1. 1 ; brhati'bhyas, viii.3.19; add padvdtibhyas, vi.59.6. 

C. Dative: apart' bhyas, i.32 .1 3. 

Dative and Ablative Plueal Neuter. 

A.B.C. The Rik has no example. In AV. ii.33.6, BR. suggest 
asthdbhyas for asthibhyas (dsthi), Abl. 

Genitive Plural Masculine. 

The oxytone stems of series A and B accent the case-ending in 
the m. and f. (and n.) of the genitive plural. This seems anoma- 
lous, since it happens although the stem and ending are separated 
by the consonant n. Cf. urmi-d' , Urmi-n-d (p. 367). 

A. I. Oxytones. There are 80 forms (from 11 stems). 1. 
kavind'm, 12; girind'm, 3; carshanind'm, vi.22.1 ; panind'm, 3; 
pathind'm, v.1.11 ; mathind'm ?, Val. 5.8; rayindm, ,31 times, 
and in vi.1.8; sanind'm. Besides these the AV.'has: kavind'm, 
vi.47.3: vii.22.1: xiv.1.53 ; girind'm, x.4.14; devajdmind'm, vi. 
46.2 : xvi.5.6 ; 8.6 ; pathind'm, ix.5.19 ; manind'm, xix.31.11 ; 
rayind'm, iii.10.5 : vi.62.2 : vii.40.2 ; 79.4; 80.3 (cf. Rik) ; 109.6: 
x.9.27: xi.1.34: xvi.3.1; 4.1; vapind'm, MSS. xix.24.6. Compare 
the numerals, trind'm, tisrnd'm, caturnd'm, saptdnd'm, ?iavdnd'm, 
dapdnd'm, saptatind'm, an'd navatind'm. 




VOL. x. 54 



398 0. R. Lanman, [i. and 2-stems. 

3. Resolution seems necessary in agnindam, x.78.3; kavindam, 
ix.64.10; rayindam, vi.45. 19; vindam, i.25.7. 

II. Barytones. There are 48 forms (from 17 stems). \. dtrindm, 
2 ; dvindrn ; dhindm, 2 ; r'shindm, 10 times, and i.84.2 ; gan- 
dhd'rindm; ptieindm; sdkhindm; havirmdthindm. 

2. The word stands at the end of a catalectio pdda of seven 
syllables, where Gr. reads -aam : dvindm, 2; r'shindm, v.66.5 ; 
tuvishvdnindm (prd adhvark is a refrain-like pad a) ; dMmindm, 
2 ; mahishvdnindm ; vdnaspdtindm ; v'imahindm ; vidtindm ; 
p&cindm ; sdkhindm, L30.ll: iv.31.3: vii.32.25 ; hdrindm, iv. 
48.5; and viii.24.17 ; 33.12; 46.1. In viii.24.23, cardnindm is at 
the end of a catalectio jagati (i. e. trishtubh) pdda. 

3. Resolution seems necessary in r'shinaam, v.65.6 : x.22.1 ; 
mu-ntnaam, viii.17.14 ; sdkhinaam, v.64.5 ; hdrinaam, v.33.2 : 
viii.24.14; 25.23: x.23.1; and ix.105.5. 

B. In viii.20.8, we may pronounce sdbharinaam, or -dm. 

C. Here belong ahi'ndm, x.139.6 ; rathi'ndm (Gv.,-aam), i. 11. 1 : 
viii.45.7 ; nadi'naam, v.74.2. The accent of atasi'ndm, viii.3.13, 
would seem to refer it to a C-stem, atasi' ?. 

Genitive Plural Feminine. 

A. I. Oxytones. There are 91 forms (from 13 stems). 1. 
krshtind'm, 9 times, and vii.26.5 ; kshitind'm, 8; carshanind'm, 
12 (m. in 463, 1) ; jdmtnd'm ; durmatind'm, i.129.86, 116/ dhau- 
tind'm; matind'm, 14 times, and iii.49.3 ; rayind'm, i.68.7 (usually 
m.) ; sumatind'm. I add from AV., grshtind'm, ii. 13.3 : xix.24.5; 
carshanind'm, xiii. 1.38: matind'm, xiii.1.33; 3.19: xviii.3.63. 

2. The word stands at the end of a catalectic pdda of 7 sylla- 
bles, where Gr. reads -aam: kavind'm, x.22.10; kshitind'm, 5; 
carshanind'm, 14, and see masculines: in x. 126.6, dti dvishah is a 
refrain; durmatind'm, Ll29.Sc, lie : viii.46.19 ; matind'm, 7; su- 
kshitind'm,' sumatind'm, 3 ; sushtutind'm, 2. 

3. Resolution seems necessary in krshtindam, vi.45. 16 : (not 
in 542,5:) viii.57.7 ; pusktlndam, x.26.7 ; matindam, x.26.4 (and 
not in 283, 3). 

II. Barytones. There are 7 forms (from 5 stems). 1. jdnindm, 
3 ; mdricindm. 2. At the end of a catalectic pdda of seven sylla- 
bles, where Gr. reads -aam : ■dpastutindm ; nirrlindm; p&cindm. 

B. It seems to be the rule for the Rik that even the oxy- 
tone stems ending in long % shift the accent to the case-suffix in 
the G.p. 

I. Oxytones. There are 11 forms (from 6 stems). 1. yatind'm, 
i.113.8 bis; 124.2; bahvind'm, i.95.4 : vi.75.5 ; bhanjatind'm, 
x.103.8 ; bhdtind'm, i. 113.15 : iv.13.1 ; bhuiijatind'm, x.89.17 ; 
mahind'm, viii.19.31 (SV. ii.1173, mahi'ndm — but not " richtiger," 
as Gr. says). We need not refer strind'm, AV. vii.13.1 : viii.6.13, 
to C, as a transition-form, since its accent is regular for B. 

Exceptions. There is no genuine exception for which the Eik does not show a 
counter-example with accented case-ending. (1.) In iii.1.12, mahi'ndm occurs with 
unconformable accent; the first part of the hymn, 1-14, is of mystical character. 



G.p.f.] Noun- Inflection in the Veda. 399 

The AV. has ndrdfansi'ndm, xv.6.4, and rdthajiteyi'ndm, vi.130.1. (2.) The word 
stands at the end of a eatalectic pdda, where Gr. reads -aam. Yati'ndm occurs 
in i.168.6 ; but verse 6, the last, belongs to the preceding five neither in sense nor 
metre. Gr. thinks it may owe its origin to a gloss. Mdhi'ndm occurs in x.134.1 
— a hymn whose style marks it as late — and in ix.102.1. Pdda, b of i.36.1 reads 
vifd'm devayati'ndm. I suspect that this is a mere metrical substitute for deva- 
ydntlndm. The law with regard to the place of the accent, p. 361 top, holds good 
in the formation of feminine participles, so that the stem must be either devayatt' 
or devaydntt. It should, like all other feminine participles of denominatives, be 
the latter (devaydntindm : cf. i.77.3); but since the metre requires a short syllable 
in the fifth place, the n is dropped and the accent shifted accordingly. For 
kant'ndm, see p. 364. (3.) Resolution seems necessary in maM'naam, v.45.3. 

II. Barytones. There are 62 forms (from 28 stems). 1. dtish- 
thantindm; drushindm; iytishindm, 3 ; dshadhindm, 6 ; gdmat- 
indm ; jdyantindm ; tr'iarushindm ; dalvindm, 2 ; ndvyasindm 
(mardtdm !) ; padvutindm; md'nushindm, 6; revdtindm; vatsin- 
indin; pdcindm; pdpvatindm, 7 ; sr'tvarindm. 

2. The word stands at the end of a eatalectic pdda of 7 sylla- 
bles, where Gr. reads -aam : aprnanmdyindtn • ddatindm, viii.58.2 ; 
A'rndvatindm; dshadhindm, 3 ; ndvyasindm (m.) ; pr'shatindm, 
2 ; md'nushindm, 2 ; ydtumdtindm,, 2 ; yoyuvatmdm, viii.58.2 ; 
rdpushindrn ; vihtitmatindm; vavarji&shindm; pdcindm,3; pdpvat- 
indm, 3 : and viii. 17.14, at end of trishtubh; piprinindm • sani- 
shydntindm. 

3. Resolution seems necessary in pyd'vinaam, viii. 46. 22. 

C. 1. Root-words: (b) dhiyd'm, v. 44. 13: — (c) dhind'm, 7; 
prind'm; hiranyavi'ndm, viii. 54. 10. 2. (c) aruni'ndm; nadi'ndm, 
14: -dm, Gr. -aam, 5; purushi'ndm, Gr. -aam,' svarfndm,x.ti8.7, 
see p. 369. Krimi'ndm — AV. v.23.13, sdrveshdrh ca kr'tmindm, 
sdrvdsdm ca krimi'ndm — should be added on p. 369. 

Genitive Plural Neuter. 
A.B.C. There is no example. 

Locative Plural Masculine. 

A. There are 37 forms (from 16 stems). These are: dkavi ; 
o.ffni, 3 ; afiji, 2 ; dsamdti; dj'i, 8 ; dpi; r'shi; khddl; gin; nidhi; 
pathi ; pavi; yoni, 5 ; raprru; pubhri; sUri, 8. Example : 
agn'ishu. For B and C there is no example. 

Locative Plural Feminine. 

A. There are 60 forms (from 21 stems — all but 3 in -ti) : as, Uti- 
sh'u. The stems are : abhlmdti; ishti; Ht'i; rshti; krshti, 6 ; kshit't, 
3; gdvishti, 6 ; dwishti, 11; deodhitti; pdrishti; pushti; prdniti, 
2 ; prdtHrti; prdyukti; yd'mahtUi, 1 ; rati, 3 ; vmshti, 13 times, 
and i.171.5 ; sdt'i; further, avyathl; jdml; gubhr'i, v.34.8. 

Transition from B: nd'rishu, x. 86. 11a. The * is metrically 
favored as penultimate in a pdda of 8 syllables. 

B. There are 64 forms (from 30 stems) : as, yati'shu. The stems 
are: dpnasvati; dmdnushi; d'runi; yati! ; dshadhi, 19, arid x.1.2; 
garbhint; gdmati; citrini; jdgati, 2 ; tdruni; tdvishi, 2 ; ndvyasi; 



400 C. R. Lanman, [w and 

nd'hushi, 3 ; pdrushni; pdrol' ; pr'shati; prajd'vati; md'nusM, 8 ; 
mrkshini; yahvi , 2 ; rdhint, 2 ; vd'ni; vd'pi; vrjani' ; pdkvari, 
2; Qapvati: -ti', p. 368; slshdsantt; snt'hiti; hdrini; jdhati. 

Transition from A: svd'hdkrtishu, i.188.11 (? metrical?). 

(J. 1. dMstvd,^. 2. apari'shu, 3 ; aruni'shu; nadt'shu, 0. 

Transition from B : strishti, AV. vi.ll'.l : vii.90.3 : xii.1.25. 

Locative Plural Neuter. 
A.B.C. The only example is hh-d'ri-shu (d'gahsu), viii.45.34. 

Declension of ari', jdni, pati, sakhi, etc. 

The stem ari is declined thus : N.s., aris, 10 ; in i.4.6, Bollensen suggests art's 
=ary-ds (vifas); A., arim, 2; arydm, viii.33.14 (G.p. -yd'ml); D., ardye ; G., 
arias, 3 ; aryds, 35 — p. 384 ; N.p.m., aryds, 16 ; A., 7 ; N.p.f., 2, and x.27.8 bis ; 
A., 4. All the forms (except the N., A., and D s.) are irregular, especially in the 
accent, and coincide with the N. or A.s.m. of the stem aryd. See Wb. 115. 

The stem jdni, t, has the following forms : N.8., jdni (transition to B) ; G., 
jdnyus ; N.p., jdnayas, 12; A., jdnts, 2, and viii.17.7 ; -ibhis,5; -ibhyas, x.183.3; 
■indra, 3. 

Pafe' is declined regularly in the sense of 'lord,' and in composition; when 
uncompounded and meaning 'husband,' it has the forms pdtye, 8; pdtyus; pdtyau, 
2 (cf. p. 388). The regular forms are : pate, 35 ; pdtis, 72 ; -im, 49 ; pdtid; -yd, 2 ; 
(for pdlind, pdtaye, pdtes, cf_. p. 382 ;) pati, 33 ; -ayas, 16 ; -ibhyas, S. 

From the stem sdkhdy come the forms: sdkhd, 76; -dyam, 27 ; -dyd, 6; -dyau; 
-dyas, 65. The others come from sakhi: sakhe, 11; sdkhid, 5; -^a, 3; -j/e, 10; 
-^«s, Ab. ; -yus, 7; -jms, 2; -fa, 5; -i&Ais, 15; -ibhyas, D., 22; -ibhyas, 4; fa<Jm, 
4; -inaam, v. 64.5. 

In like manner are declined most of the compounds of sakhi. The following 
are found in the N.s.m. : asmdt-; indra-; kdnva- ; go-shakhd; nrvdt- ; mat-; 
marut-, 4 (and as fern., vii.96.2: x.86.9); sushakhd', 2. Besides these, are found: 
gosakhdyam ; sushakhd'yas, 2. Pour compounds show transition to the a-declen- 
sion: drdvaydt-sakha-m, x.39.10e; ydvayat-sakhd-s, x.26.5d; pataydt (sc. -sakham) 
mandaydt-sakha-m, i.4.7c. For the omission, cf. the German Ein- und Ansgany. 
I suspect the metre has something to do with these transitions ; and perhaps the 
tradition once had frdyavdtsakhas at the end of viii.46.12a, where our text has 
■khd. 

The peculiar form in -yus is common "to jdni, pati, and sakhi. Bopp, Vgl. Gr. s 
§198, considers pdty-au and sdkhy-au as organic forms. I should regard them 
rather as due to false analogy, -au being apprehended as a case-ending. 

In the AT., x.8.39, we have the N.s.m., ekapatni-s. This form, like hiranyavdci-s, 
RV. vii.97.7, belongs to a category that I have not distinctly set up, namely, the 
masculine compounds of feminine stems in i. 



STEMS IN U AND ti. 

The declensional system of the u and Osteins is closely analo- 
gous to that of the stems in i and I (p. 365). There is, however, 
owe important difference. The forms of the i and «-stems show 
that there were three distinct methods of declension, by means 
of three series of endings, A, B, and G. The forms of the 
stems in u and H show that in the Vedic period their declensional 
resources were confined entirely to the series A and C. The 
few sporadic instances of B-forms from w-stems only confirm this 
statement. The forms which are even preferred by the later 



Series A. 


£ 


Series C. 


o as 


&'s 


A' as 


A ns 


A' am 


A' as 


u bhis 


A' & 


A' bhis 


u bhyas 


A' e 


A' bhyas 


An A'm 


A' as 


A'n Am 


u shu 


A'i 

u 


A' shu 


u u 


A' A 


u bhyAm 




A' bhyAm 


u 6s 




A' os 



■^-sterns.] Noun-Inflection in the Veda. 401 

grammar — as, dhenvai, dhenvd's, dhennd'm — are in reality transi- 
tions to the *-declension, and do not belong to the w-declension at 
all. As stated at the eud of p. 366, the apparent case-suffixes of 
the Osteins were -4, -ai, -ds, -dm, etc. ; and after this analogy the 
forms dhenvai, etc., were made. The originality of such a form 
as hdnv-ds (Schleicher, Gomp.* p. 538 §252) must be denied. 
The two series, mutatis mutandis, are as follows : 

u s 
u m 
u A' 
o e 
o s 
o i 



Series A was applied to the stems in short u, including many 
masculines, but few neuters, and very few feminines. It was 
applied without distinction of gender, save that the neuters of 
course had in the N. and A., no ending in the s., 4 in the d., and -i 
in the p. The later grammar has rules confining the "inserted w" 
to the neuter in certain cases ; but in the Veda we have neuters 
without n (as, vtdau, mddhau) and masculines with n (as, 
cd'runas). Here belong, moreover, a number of stems in ■&, the 
character of whose final is not always entirely clear, but which are 
treated declensionally as if the u were suffixal. 

In some, etymological evidence shows that the u is not radical. Such are : 
abhi-jnu, mitd-jnu, dru, su-dru, pMu-dru, snu, ghrtd-snu (cf. jd'nu, dd'ru, sd'nu) ; 
khara-jru (root 2jar) ; d-gru (gir) ; puru-kshu, Jcshu (ghas) ; d-psu (bhds) ; -psa 
(bhAs), with arund-, ahruta-, rta-, prushita-, vifvd-, vifod'-, vr'sha- ; sishnu (san) ; 
pipru (par). Gu, in the compounds of -gu with d-, pushti-, pr'frd-, bhA'ri-, saptd-, 
and in su-gu, is a peculiar modification of go, ' cow.' We might see its parallel in 
rdhdd-ri, p. 384, if the metre did not make the assumption of a stem -ri doubtful. 
In other cases, the final a (or a + nasal) of a root is dropped and replaced by the 
suffix u: thus, ^(genitive, yds), sva-yu, fubhark-yu (cf. fubharh-ya' ) ; anu-shthu, 
su-shthu (sthd ; but cf. sthA-rd, sthav-ira) ; ddhri-gu, vanar-gu, fdci-gu (gam) ; A-khu 
(khan, khd-ta). 

Some words whose final element is a root ending in a short vowel are declined 
entirely as if the u were a suffix. Such are : dyu, ddyu, dbhidyu, ddfadyu; didyu; 
dprdyu (yu, yuyoti) ; dsmrta-dhru (cf. varuna-dhru-t) ; prthu-shtu (stu, ' congloban') ; 
su-shtu (stauti). In mitd-dru, the root is used substantively as part of a bahuvrihi; 
and so possibly in sd-dru, AY. xv.7.1. In hari-dru, Rik x.94.12, the accent is 
probably wrong ; read haridrdvas. 

The compounds of bhu ought to be declined according to C ; but even in the 
Veda there are many forms from stems in short u. In the enumerations, I have 
entered them as transition-forms (C to A.) ; but perhaps the general statement, p. 
403, would have sufficed. 

Series C was applied to oxytone stems in long -A', and their 
sometimes barytone compounds. Here belong 

1. Stems whose final element is a root: as, nabho-jH', vira-sd 1 '; 
monosyllabic substantives and their adjective compounds: as, jil', 



402 G R. Lanman, \u and 

sii', m. ; <W, bhd', bhr-d', sru', f. ; subhu' ', m.f. Two words whose 
final element is a root ending (according to the dictionaries) in H 
are declined as if the final were U. For such forms we ought 
probably to set up stems in long vb as follows : raghu-dru' 
(drdndud, iv.4. 1 ) ; ghrta-swti' (cf. ghrta-snd' ', prd-snautu, nods). 
Snd: snil : : dd : dd. See Gr. 623, and cf. gd and j6gd. To 
these we may perhaps add mitra-krd' ; cf. krd-rd. The sww of 
vrdhaswd is a suffix — -Lindner, p. 112. 

2. A few oxytone f. substantives (a) corresponding to barytone 
masculines: agrd' (dgru), 'maid;' kadrd' (kddru), ' soma-vessel ;' 
gugguld! (gtiggulu), an Apsaras ; jatH' (jdtu, n.), 'bat;' dhamd' 
(dhdnu), 'sand-bank;' prddtcd' (pr'ddku), 'serpent;' pvaf.rH! 
(pvdpura), ' mother-in-law ;' likewise the adjective mad/itf! (tn ddhu) , 
' sweet.' (b) To several others there is no corresponding m. : 
kuhd', ' new-moon ;' catwd', ' bowl ;' juhti,', ' tongue ;' juh'd', ' sacri- 
ficial ladle ;' tan-Ci! ', ' body ;' prajanH' ; vadh-d', ' bride ;' vifil' ?, 
' nest.' 

3. The feminine of a large number of stems of very varied 
character (p. 367) is formed by long i, and the largest part of the 
stems in i are feminines of this description. Long $ is not used 
in forming the feminine, except that of &-stems. And of these, 
many remain unchanged in the f. (cf. subdhus, m., viii.17.8 ; f., 
ii.32.7), while others take long i. The stem tanu forms its f. in 
all three ways : tawd- ', tayvd, and tanvi. (Cf. the double feminines 
prabhtt', prabhvi' ; madhd' , mddhvi; vibhH', vibhm ; bibhats-d, 
-8-d'.) For the Veda, it seems necessary to set up the following 
feminine stems in 'd' (from the Rik and Atharvan), corresponding 
to oxytone masculines in Hi : ahhoyd! ; asitajn-d'; tawd', adj. ; 
durhandyd'; pataydl-d' ; pdrayishn-d'; prpaiidyd' ; mandrayd'; 
mahiyd.' ; mumukshd' ; sandyd' '; sixdayitwd' (of these the m. does 
not actually occur in the Rik) ; aghdyd' ; apasyd'; abhidips-d' ; 
avasyd' ; dyd! ; udanyu' ; caranyd' ; carishnti.'; jighatsti! ' ; didhi- 
shil' ; duvasyd' ; dravitwd' ; panasyd' ; prtandyd' ; babhrH'; 
blbhatsU' ; makhasyd 1 '; vacasyd'; vipanyd'; pundhyd' • sudrd' • 
further, nabhan-d' ; nrt-d! '; Gungil' ; Saranyu; Kama-dyd'. The 
last five are substantives. The -d of kama-dyd' is rather char- 
acteristic of the f. than radical ; but a m. kama-dyd does not 
occur. In like manner mitra-krd' might be taken as f. of mitra-kr-d. 
The AV. has kid'mbil, corresponding to kiyd'mbii, n., of the Rik. 

Finally, several masculines (other than those included under 
1.) belong to C: prdpH' ; maksh'd' ; krkaddpd'; dtaptatanu; 
sdrvatawd. 

The final thematic H is sometimes " split" into uv before the 
vocalic endings of C. This occurs regularly (1) in monosyllabic 
substantives always : as, juvas, bhvva; (2) in compounds whose 
last member is a verbal root : as, nabho-jtivas, mayo-bhAvas; (3) in 
feminine adjectives in H' preceded by y, to avoid the combination 
yv : as, apasydvas; and so with agrd', kadrd', blbhatsd' . The 
forms of compounds of verbal roots in t' are written with iy only 
when a double consonant precedes {brahma-prtyam, gdthd-mam, 



M-stems.] Noun- Inflection in the Veda. 403 

p. 369). This restriction does not hold for the stems in radical it'. 
The Rik has uv very often after a single consonant : as, vasu- 
juvam. This happens in the great majority of instances at the 
end of a pdda, and possibly the syllable uv was written in order 
to fill out the metre for the eye. Of course, vibhvdtn has just as 
many syllables as dbhuvam. The stems whose forms are written 
as we should expect — with u rather than uv — are in fact the 
exceptions, and are: ghrtapu', vishnd-pu' / vi-bhu', su-bhu' ; 
surdrfu' ' ; d-sd, nava-su', pra-sii', vi$va-su". The TS. appears to 
have no fixed rule with regard to the use of uv. Cf. tanuvam, 
i.5.5 4 ; indravdyuvos, iii.1.6"; bdhuvds. i.7.13 2 ; uruvos, v.5.9 2 ; 
suvar, always; but svasti. See Ind. Stud, xiii.105,106. 

In the Rik, the distinctions between the u and w-declensions 
(A and C) are for the most part strictly maintained, except in the 
compounds of bhu. The transitions to B are merely sporadic. 

I. Transitions from A. 

A to B. The Rik has only three forms that fall under this 
category: ishvai, vi. 75.15 ; 'tshvds, x. 18.14 ; suvd'stvds, viii. 19.37. 
All these are plainly verses of late date. The AV. has urvdrud's. 

A to C. In viii.46.6, we have (indram) dbhiruam. In x.28.8, 
perhaps sudruam is to be taken as f. Although a root-word, 
dprdyu is properly declined according to A (p. 401) ; but in i.89.1, 
we have dprdyuvas. The masculines madhydyuvas, mitrdyuvas, 
Qramayuvas (N.p.), and madhuyuvd (V.d.) are hardly to be 
accounted for, except as transition-forms. We need not try to 
explain the grammatical monstrosity pag,vd! (N.d.m., accent !), 
x.106.3. The character of the hymn is such that none of its 
anomalies can surprise us. The TS. has hdrvdbhydm. 

II. Transitions from C. 

to A. This is the most important category, even in the 
Veda. It becomes still more so in the later language, where 
stems ending originally in long u are replaced by others in short 
u. Cf. BR., s.v. prabMi. In the Rik, about ten stems ending 
originally in u show secondary forms (about 35) from stems in 
u. In ix.86.1, we have dhi-jdvas (dhi-ju'). Gr. derives sabar-dhu-m 
(dhenum), x.61.17, from dhu ; cf. dhu-noti and dhii-noti. The 
compounds of bhu pass into the ^-declension with especial fre- 
quency. Thus we have: dprabhu, -us; dbhu (from d-bhu' ?), 
-■dm 2 ; purubhu, -u, V.d.m. ; prabhu (AV. xiii.4.47, prabhu'-s), 
-us 4, -<im^-6s; mayobhu, -us 4, und, -u' (N.p.n.), -u'n; vibhu, -us 
8, -4m, -dvas (for Vibhubhis, see IST.p.m.) ; pambhu, -us 2, -u 
(V.d.m.), -u' (A.d.m.). Perhaps the short u of (vikshu) dyu'shu, 
i.58.3, is due to the metre, since it is the penultimate of a jagati- 
pdda. For bibhatsu's (mdtd 1 ), see p. 402 med. 

The Atharvan has: abhibhfi'm, x.6.29; purubM', iv.25.1, where there is good 
reason for restoring the true C-form, -bhuvd (see N.d.m.); mpu'nd, xii.2.11 (cf. 
pavitrena su-p&'-d, from su-puf, VS.i.3) ; for dushtano, see V.s.m. The VS. has 
from tidbhti', (pdrman) udbhau, xv.l. 



404 C. R. Lanman, [u and 

C to B. The only instances in the Rik are pvaprud'm, x. 85.46 
(a late verse), sud'm (? see L.s.f.), and dravitnu-d' {dhiyd'). The 
last I regard as a transition-form on account of its accent : the 
C-form would be -nti-d. On the other hand, it is better to refer 
mehatnu-d' to a stem -nib' (not -nu'). Tan/A' as a dual (x.183.2) 
would be unparalleled ; see L.s.f. 

The Atharvan shows a good beginning in this direction, and contains the fol- 
lowing B-forms : kuh&'m, tanii'm, vndh&'m ; vadhvai, fvafritai ; punarbhuvds ; 
prddkud's, fvafrud's ; tanudm ! ; agruvai and rajjvdm. The VS. has a-su'm T for 
a-su'-am. 

III. Transitions to the N-Declension. 

In explaining the case-forms with internal n from i-stems (p. 
373) it was natural to point to the analogy of the frequent in- 
stems ; but since there is no class of stems ending in un, grave 
exception may be taken against the term "w-declension." I have 
used it, however, for the sake of convenience and brevity. 

As with the i and ?-stems, this " transition" is universal in the 
G.p. of A and C, the only exception being jdguvdm. The n is 
also seen in series A in the I.s.m. (108 forms, 31 stems), in the 
G.s.m. (cd'runas and drdnas), in the I.s.n. (39 forms, 15 stems), 
D. (mddhune), Ab. (mddhunus, sd'nunas), G. (2'7 forms, 4 stems), 
L. (d'yuni, sd'nuni), and in the N.A.p. (127 forms, 14 stems). 

Here Benfey puts abhini-narn, despite the accent ( Vedica, p. 
123— but see" 131), VS. vi.17 (= AV. vii.89.3, abhi'runam). In 
ix.65.30, d' ray'vm d! sucetunam (vrntmahe), one must see a stem 
sucetuna = sucetand, or change the reading to suceti'md (adv.), 
or take the form as A.s.m. of sucetti. 

The following synopses exhibit the terminations of the 
inflectional forms : 

Series A. Masculine. 

Singular : N., -us, -w'/ A., -urn • I., -ud, -vd, -und ; D., -ue, -ve, 
-ave; Ab., -vas, -os; G., -uas?, -vas, -unas, -os, -av-asf; L., -av-i, 
-au, -d?; V., -o. 

Dual: N.A.V., -fl {iti), -it; I.D.Ab., -ubhydm; G.L., -uos. 

Plural : N.V., -avas, -uas, -vas, -dvas? ; A., -Un, -iinp, -Unr, 
-vas; I., -ubhis; D. Ab., -ubhyas ; G., -■O.ndrn, -unaam ; L., -ushu. 

Feminine. 

Singular: N., -us; A.,-um; I.,-ud,(-u-y-d',&dx.,)-vd; D.,-ave; 
Ab.G., -os; L., -au; V., -o. 

Dual : N. A.V., -A (iti) ; L., -uos. 

Plural: N.V., -avas, -vas; A., -Us, -uas; I., -ubhis; Ab., -ubhyas; 
G., -ilndm, -Unaam; L., -ushu. 

Neuter. 

Singular: N.A., -u, -it; I., -vd, -uf, -und; D., -ve, -une, -ave; 
Ab.G., -uas, -vas, -unas, -os; L., -un-i, -av-i, -o, -au; V., -u. 
Dual: N.A., -vi (iti). 



w-stems.] 



Noun-Inflectwn in the Veda. 



405 



Plural: N.A., -4, -fb, -tint; I., -ubhis; G., -■dndm, -tinaam; L., 
-ushu. 

Series C. Masculine. 

Singular : N., -$«/ A., -nam; I., -?<$/ D., -ue; G., -was. 
Dual: N.A.V., -ud. 

Plural : N.V., -wassy A., -uas, -'tis; I., -ilbhis; G., -udm, -tindm; 
L., -Ushu. 

Feminine. 

Singular: N., -$s; A., -nam, -vain?; I., -wd, -#,?/ D., -we/ 
Ab.G., -was/ L., -ui, -ui, -H (iti) ; V., -u. 

Dual: N.A., -ud; G.L., -uos, -vos?. 

Plural: N., -uas, -us, -vas; A, -was, -vas, -Us?; I., -Ubhis; D., 
-ubhyas; G., -Undm, -tinaam; L., -ilshu. 

Neuter. 

Singular: N.A., -m; I., -ud. 

Oxytone stems, when not strengthened by vowel-change or 
nasal, shift the accent to the ending in series A, in the I.s., in 
the Ab.G.s., in the G.L.d., and always in the G.p. In C, on 
the contrary, the accent never leaves the thematic vowel, 
except in monosyllabic stems. Sporadic exceptions will be 
noticed as they come up. 

Several A-stems have forms analogous to those of ari ; i.e., the vocalic endings 
are affixed directly to the " unstrengthened" stem : as, papu-ds (=pac&'n). These 
stems are: pacu; pitu; kr'tu; krdtu, fat&kratu; madhu ; vdsu; Vibhu' ; fipu; 
sahdsrabdhu. 

By way of illustrating the wealth of declensional resources in the Veda, a few 
equivalent forms may be cited: krdtud, -vd, -und ; krdive, -am; krdtvas, -os ; 
cd'runas, cd'ros; drunas, dros ; druhydvi, druhyau; pafvd', paound (accent, p. 367 
top); pafvds, pac&'n ; fipve, fipave; (ifvas, fifos; tanuam, tanu'm; vadhAam, vadh- 
u'm ; camui, camui, camu' ; uru, ur&' ; puru, pwu,' ; m&dhuas, -vas, -unas, -os; 
vasvas, -unas, -os : sd'nunas, -os ; sd'nam, sd'no, sd'nuni, sd'nau ; abhilhuvam, 
■bhu'm; mayobhu's, -bhii's ; -bhuvd, -bhu'nd; -bhuvas, -bhu'n; vibb.ua/rn, -bTiirni ; 
vibhuas, vibhdvas ; vibhuas, vibhvi's ; cambh&'s, cambhu's; etc., etc. 

Nominative Singular Masculine. 

A. The ending -s is affixed directly to the stem. The form 
occurs 886 times (from 250 stems). Examples are: anpus, 13; 
asmayus, 16; Indus, 64; rbhns, 21; ketus, 23 ; krdtus, 14; cd'rus, 
16; devayus, 15; rndnus, 15; vdsus, 25; vdyus, 20; vipvd'yus, 
18; vlshnus, 34; smdhus, 17; sukrdtus, 32; sil.nus, 18; so ddhri- 
gus, 3; khardjrus; purukshus, 5; mitddrus, 2; yx'ts, viii.18.13 — 
BR., dvayus; svayus, 2. In viii.66.7, tshus is m. 

Transitions from C : dprabhus ; prabhtis, 4 ; mayobhus, 4 ; 
vibMts, 8 ; pambhfts, 2. 

In vii.86.3, prche tdd eno varuna didr'/cshti'po, p. -shu \ upo, we 

have elision and crasis. Say., chdndasah sulopah : drashtum 

ichann aham ; but since desiderative adjectives are oxytone 

(Lindner, p. 62), perhaps didr'kshu is an adverb with recessive 

vol. x. 55 



406 C. R. Lanman, [u and w-stems. 

accent. So in ix.96.15, urv Iva gattih suydmo nd vddhd: Say., 
supdrh sulug iti sor luk. Gr., urAr va. 

In vii.60.3, visarga seems wrongly added before $■ in dhd'rndni 
mitrdvarund yuvd'kuh \ sdrn. BR. suggest yuvd'ku, as p.n. with 
dhd'rndni. Cf. sdsni | sdrn, p. 377 ad fin. 

C. 1 . Here belong 24 forms (from 9 stems) : -bhu's with abhi-, 
6 ; pari-, 9 ; puro- ; mayo, -3 ; saed- ; svayam- ; ekadyA's (root 
Idiv, dyu-td) ; s-A's, i.146.5 ; vi$vd$ambhils; and from the Athar- 
van, udapil's, madhup-A's, vdtapH's, xviii.3.37; prabfiU's, xiii.4.47; 
viblvA's, vii.21.1 : xiii.4.47: xv.15.6; pambh-A's, xix.46.6 ; subh-A's, 
xiii.4.52. 

2. PrdpA's, i.40.1 (ER., prdpus) ; dtaptatawAs; sdrvatamus, AV . 
v.6.11-14: xi.3.32-49 bis: <>t. Br. iv.6.1 1 : xi.1.8". 

Nominative Singular Feminine. 

A. There are 73 forms (from 31 stems). 

Enumeration : arundpsus, 2 : aharhyus ; dbharddvasus ; ishus, 2 ; krmmus ; gdtus, 
2 ; caranyus ; cd'rus, 2 : jasws ; jivd'tus, 2 ; tridhd'tus ; 1 dd'nus ; Idd'nus ; dhiyd'- 
vasus; dhenus, 22; pdrpus; btbJtatsus, see p. 402 med. ; bhujyus ; rajjvs ; r'upat- 
papus ; vasitytts, 3 ; vastus ; f&rus, 4- ; pravasyus ; saptddhdius ; sardyus. 2 ; sindhus, 
11; subdhus; sumadanpus; suprotus; sudbhipus. 

In i.186.4 and vii.2.6, visarga seems wrongly written ; for su- 
dugheva dhen/th, Gr. proposes -e va dhenti'. 

C. Here belong : 1. bhU's, 2 ; punarbh-A's; mayobhU's; s-A's, 
i.32.9; yamasAfs ; rahasil's; virasii's; sushti's; from the AV., 
abhibh-A's, ii.27.1 ; prabhtt's, ix.4.2 ; fambh-A's, xiv.2.26 ; prasil's, 
iii.23.4. 

2. Tan/A's, 8 (AV., 7 : as, ii. 13.4) ; vadM's, 4 (AV., 5: as, 
i.14.2) ; fvaprd's ; from the AV., kuhU's, vii.47.2 ; gvggvM's, 
iv.37.3 ; juh-d's, xviii.4.5 ; dhan/A's, i.17.4; prddkfl's, v.18.3,15 ; 
madhti's, vii.56.2 ; from the VS.,jat-A's, xxiv.25,36. 

3. GungH's ; nrt-A's; gundhyus; saranyil's; from the AV., 
kid'rnb'ds, xviii.3.6 (cf. Rik x.16.13) ; asita-jfM' s, xii.1.21 ; pata- 
ydlil's, vii. 115.2. 

Nominative and Accusative Singular Neuter. 

A. 1. The form has no case-ending. It occurs 413 times (from 
68 stems). Examples are : urn, 51 : and iv.43.5, as adv. ; cd'rx, 
27 ; tridhd'tu, 13 ; pdpu, iii.53.23 ; VS. xxiii.30 ; puru, 13; prthu, 
15; mddhu, 76; vdsu, 77; sd'nu, 18; further, abhijwi, 5; ksh-A ; 
vi$vdpsu; sushthA; dprdyu. 

For didr'kshu, vii.86.3, and uru, ix.96.15, see m. In iii.49.2, 
the metre requires d'yu, text d'yus. In AV. viii.2.28, the sense 
requires pdrayishryA — text -%'is (-$' r-) ; in iii.19.1, jishryA — MSS. 
-■As; in xx.l35.12,'5aM — MSS., -■Ah; and in xix.30.1, jard'mrtyu 
— MSS. and ed. -uh p-. 

2. The final -u suffers metrical protraction in several words : 
■ur-A', vi.47.14 ; purfi! , 12 times, given by Gr. : as, x.94.5 ; rn'tth-A,, 2. 
The pada has in every instance -ii. Cf. Prat, vii.31 : vii.9,19 : ix.3. 



N.A.s.n.] Noun-Inflection in Ike Veda. 407 

C. Since the forms for A and C are coincident, it is sometimes 
doubtful whether a word is to be referred to a stem in -d, or to a 
transition-stem in -■&. Here may be put : dbhu; prdbhd, 2, see 
BR. ; mayobhu, 4 ; vibhu, 4 ; pambhu, 2 (the foregoing have other 
&-forms) ; subhu (annum), ii.35.7; from the AV., (pdrma) udbhu, 
ix.2.16; vibhu, prabhu, xx. 135.9 ; pambhu, x. 1.9 ; and also raghu- 
dru (cakrdm), Rik x.61.16, since the only other forms of this word 
require a stem -drd' (p. 402 top). 

Accusative Singular Masculine. 

A. The form terminates in -urn, and occurs 736 times (from 179 
stems). Examples are : anpum, 17 ; dpiim, 20 ; indum, 27; urum, 
18; ketum, 34; brdtwm, 55; \gdtum, 33; cd'rum, 14 ; tdntum, 
15 ; ddsyum, 15 ; papum, 15 ; bhdnum, 15 ; bhujyum, 15 ; man- 
yum, 18 ; vdyum, 18 ; pdtrum, 15 ; pipum, 26 ; shtdhum, 22 ; 
sunum, 17; further, ddyum; ddhri-gum, 4; abhidyum; d-khimi; 
ddpadyum, 2 ; didyum, 6 ; piprum, 7 ; purukshum, 7 ; pr'pnigum; 
prushitdpsum; vipv&pmm; saptdgum. 

Transitions from C: dbhum, 2 ; prabMim; vibhum; abhibhum, 
AV. x.6.29. 

C. Here belong : 1. (a) vibhuam, 4 ; vishnd-puam, 3 ; subhuam. 
(b) dbhuvam, 2 ; kapojuvam / manojuvam, 2 ; mayobhuvam, 5 ; 
vastijuvam; vipvdpambhuvam; pambhuvam, 2 ; sacdbhuvam, 2 ; 
sudbhuvarn, 2 (these nine words are in every case at the end of 
a pdda of 8 or of 12 syllables, except manojtivam, x.81.7) ; abhi- 
bhuvam, AV. ix.5.36 ; pambhuvam, x.6.15,17. 

2. Krkaddpuam may be taken as m. with sdrvam, i.29.7. 

Transitions from A : dbhiruam, viii.46.6 ; and perhaps sudruam, 
x.28.8 (or f.). 

Transition to the M-declension : sucetunam (?). See p. 408 med. 

Accusative Singular Feminine. 

A. The form occurs 50 times (from 15 stems). 

Enumeration : ddhenwm ; arunapsum ; ishum, 2 ■ krumv/m ; cd'rum, 2 ; jaswm ; 
jigatnum; jiraddnum; jiod'tum, 3 ; dhanum; dhenum, 27; bhujyUm; vasuyum; 
(arum, 3 ; sindhum, 4. 

Transition from C : sabar-dhd-m, x.61.17; of. p. 403 ad fin. 

C. Here belong : 1. (a) a-suam, 2 ; sakrt-stiam • for su-dm, 
v.58.7, see L.s.f. ; from the AV., as-dam, vii.35.3 ; vipvas-dam, 
xii.1.17. (b) vipvaj'uvam: sacdbh&vam; from Gat. Br., bhruvaru, 
iii.2.1 2 \ " 

2. Prddkuam, AV. x.4.17 (MSS. and ed. -yam/); RV.Juh&am, 
2 ; tantiam, 45 ; vadhuam; in true Atharvan verses, tantiam, 16 
times; tanvdm, xix.37.2. 

3. (d) Garishnuam (ptiram), vii\.\. 28 ; sudruam, vii. 32.20 (SV. 
i.238, -iivam): possibly, x.28.8; jighatsuam, AV. ii. 14. 1, (b) 
avasyuvam ; durhandyuvam ; makhasyuvam ; vacasyuvam ; 
pundhyuvam, 2 ; kamadyuvam. (c) tanvdm ?, as an adjective, 
viii.65.12c. 



408 C. R. Lawman, [u and w-stems. 

In only about four passages of the Bik does long thematic & become v. They 
are: x.51.26, \c: ix.96.21c: i.162.206. The fourth passage is a late one. In 
ix 96.21c, possibly the shorter grammatical form has been ousted by the longer : 
read krl'dcm camuor a' vipa pundna (text cam/ear . . p&yamdna) 1. Perhaps, 
too, we ought to accept Grassmann's suggestion and read mdrdya tanuam mame, 
viii.65.12, text indrdt pari. Cf. p. 319 top. 

Transitions to B. The case-ending of devi'-m was apparently m 
(p. 366 end). After this analogy were formed : kuhdm, AV. vii. 
47.1; vadhu'm, ix.3.24 : x.1.1 ; tanu'm, xviii.1.13,14 (the Rik, 
x.10.12, has the older form, tanuam) ; asH'm, VS. xxx.14. The 
llik has no such forms in -iX'tn. 

Instrumental Singular Masculine. 

A. The general ending is added directly to the stem : 1. 
krdtud, iv.28.3: vii.21.6. 

2. The stem-final becomes v : in the oxytones parapvd', i. 130.4, 
and papvd' , twice ; in krdtvd, 57 ; and in pipvd. 

Transition to the w-declension. This occurs in 108 forms (from 
3 1 stems) : as, krdtund. 

The stems are : ahgit ; aktu, 2 ; andfu ; abandhu ; indu ; uru ; rju ; rtu, 8 ; ketu, 
9; kratu, 12; cetu; jishnu; tridhd'tu; dru,5; dhUmaketu; dhrshnfa; papu; bhdtm, 
24 ; ma/nu ; manyu, 4 ; ripu ; vagnu, 2 ; vahatu ; vdyu, 1 ; vibhindu ; vishnu, 4 ; 
vr'shapsu; sddhu, 2; sncetu, 7; stanayitnu; snu, or n., 4. 

Since swd and dru are apparently monosyllables, we might 
expect the accentuation snund', snubhis, snushu, drund' (cf. dht- 
nd'm) ; but in reality these words are accented as if from the 
dissyllabic stems sd'nu, dd'ru, of which they are the shortened 
forms. Compare tmdn-d (p. 341 end), tmdne, tmdni (for dtmdnd, 
or dt-, etc.) ; yii'rie, yU'rias (for ydvane, -as) ; pun-as (for pudn-as) ; 
su'r-as (for suar-as) ; but stri-bMs. 

If sucetunam d', ix.65.30, stands for sucetund d' (the m being 
introduced to avoid hiatus, Gr. 1531), the case is parallel to that 
in vii. 7.2 (p. 348 end). But cf. p. 404 med. 

Transition from C : mayobhtind, Hi. 16.6. 

C. Here belong : vibhud; manojuvd; sendjuvd; sudbhH,vd. 

Instrumental Singular Feminine. 

A. Here belong: 1. mehatnud! (stem -nu', on account of accent); 
ddhenud; pdrud, 5 ; susdrtud; hdnud; ishud, AV. v.5.4 ; hdnud, 
vi.56.3. 2. panvd' (stem -nu') ; cikitvd', AV. vii.52.2 ; mddhvd, 
RV.,2 ; tshvd, AV. x.1.27. 

Transition to B : (from C) dravitnud', viii.81.15 (stem -nu\ 
x.49.9). See p. 404 top. 

Six oxytone stems in u have instrumental in -u-y-d', with 
adverbial displacement of accent : anushthuyd' ; amuyd', 7 ; 
dpuyd', 2; dhrshnuyd', 15; raghuyd' ; sddhuyd', 6; similarly 
mithuyd' from m'lihu. Of. p. 358 top. 

C. Here belong : 1. bhuvd', 2 ; punarbhuvd, AV. ix.5.28. 2. 
ljuhud, 6; 2juhud, 4 ; tan-da, adj.; tanud, subst., 27 ; vadhud; 
in true Atharvan verses, tanud, 4 ; tanvd\ 5 ; rdjjud (iii.11.8). 
3. duvasyuvd. 



I.s.f.] Noun- Inflection in the Veda. 409 

In x.24. 1 , s&rnam piba camd' sutdm, cani-A' may be I.s., for -u"d: 
' Drink with the cup the soma pressed' (cf. tapanf, p. 381 end) ; 
or L.s. : ' Drink the pressed soma in the cup.' 

Instrumental Singular Neuter. 

A. Here belongs: 2. mddhvd, 17; AV. v.27.2,3: vi.12.3. 

Possibly the crystallized case-form jdiit'- (with -bhartnan, 
-shthira) is an instrumental, for jdtud, like matt, p. .380 ; but jd'tib 
may be an adv. accusative. SuhdntH, (2) might mean ' with easy 
blow ;' but see A.d.m. and A.p.n. 

Transition to the w-declension. This occurs with 39 forms (from 
15 stems): as, urund. The stems are: uru; krdhti • ghrtdsnu; 
jard'yn; titail; trsku; tridhdltu; dd'nu; dhrshnu, 3 ; purdru; 
prthu; mddhu, 16 ; vdsu, 7 ; sd'nu, 2 ; svddu. 

Transition from C : supund, AV. xii.2.11. 

C. Here belong : mayobhuvd; (pavitrend) supu'd, VS. i.3. 

Dative Singular Masculine. 

A. Here belongs : I. sahdsra-bdhu-e, viii.45.26. 2. krdtve, 
11 ; pdpve (n. ?), 3 ; pipve; krdtve, AV. vi.19.2 : xviii.2.23. 

3. The stem-final is usually gunated and -e added : thus, dyo-e, 
dydve. This form occurs 231 times (from 66 stems). In i.61.1, 
the metre seems to demand ddhrigdve, text -dve. 

The stems are : anpu, ; aketu ; aktu ; ddhrigu ; abhidyu ; avasyu, 2 ; avishyu ; 
dyu, 10; dfu; indu,i; iyakshu; uda/nyu; rju,1; r'shibandhu ; kakdrdu; kavatnu; 
kdru,6; krtnu; ketu, 2; krdtu; kshiprd-ishu ; grhi,; ceru ; jantu ; jtrdddnu ; tdku; 
tr'tsu; ddsyu, 9; dushtdritu; dhrshnu, 3; panasyu; parimanyu ; pa$u; pdyu; 
piyatnu ; p&ru, 5 ; pedu, 5 ; prdyajyu ; babhru, 2 ; bahu ; bhdnu, 2 ; bhtru ; bhr'gu, 
2; manasyu; mdnn,36; manyu, \1 \ niddayitnu; mrtyu, 6; ydjyu,5; yddu; 
ripu, 1 ; vacasyu, 3; vdsu, 3; vdyu, 23; vishnu, 13; vidu, 2; vetasu ; fdtru; 
(dihtanu, 3; fayu, 1 ; f ifu, 2 ; sudd'nu, 6; subdndhu ; sunii, 10; svabhdnu, 2 ; 
hatnu. 

G. Here belong : vibhue; vishndpue; subhue; and, abhibhuve; 
vipvdbhuve; saedbhuve. 

Dative Singular Feminine. 

A. Here belong : jivd'tave, 7 ; dhendve; pdrave, 5. 

C. Here belongs tani'ie, RV, 18 ; AV., 6 ; -ve, AV, 7. 

Transitions to 13. The Rik lias only ishvai (from A), vi.75.15. 
The Atharvan has 5 cases, all from C : vadhuai, xiv.2.9 ; pvapruai, 
26 ; vadhva\ 73. In agruvai, vi.60.1,3, the 0, is "split" after gr 
(but not in fvayruai after pr), and so the accent is not shifted to 
the ending. Cf. punar-bhuvds, G.s.f. 

Dative Singular Neuter. 

A. Here belong : pdyve ?, 3 ; urdve; vipvd'yave. 
Transitions to the w-declension : mddhune ; kapipune, AV 
vi.138.5. 

C. For the remaining cases of the n. there are no examples. 



410 0. R. Lanman, [u and it-stems. 

Ablative Singular Masculine. 

A. 1. Here belongs only pitods, vi.20.4. 

2. The usual form in -os occurs 33 times (from 22 stems). It is 
a question whether cdkshoh (su'ryo ajdyata), x.90.13, is not a 
mere ungrammatical (or contracted ?) form like vidydt. The sense 
of AV.ii.35. 4 suggests the reading cdkshos, text cdkshus. See 
V.s.m., and stems in t radical, Ab.s.f. 

The stems are: ahhu, 7; adhvaryu; rtu; krcd'nu, ix.77.2; kshipanu; gdntu ; 
caksu ? ; tanayitnu ; ianyatu ; devayu ; dyu, 2 ; ninitsu, ■ pO/ru ; many it, 2 ; mrtyd ; 
ririkshu; vanishthu; v&yu, 2; vishnu, 3; catru-catru; sindhu; snu, 

C. There is no example. 

Ablative Singular Feminine. 

A. Here belong only : dhduos, 3 ; vdstos, 3. 
C. Here belong : 1. bhuvds ; asus'das, AV. x.10.23. 2. 

kadrdvas; taivdus, 3 : AV., 2. 

Transition to B : (from A) ishvds, x.18.14 : AV. v. 14. 12. 

Ablative Singular Neuter. 

A. 2. Mddhvas occurs twice as Ab. 3. The usual form 

occurs 7 times : uros, 4 ; sd'nos; svddds, 2. 

Transitions to the w-declension : mddhunas; sd'nunas. 

Genitive Singular Masculine. 

A. Here belong: 1. perhaps dhrshnu-ds, x. 22.3 ; ripu-ds, iv.3.13 ; 
pipru-as (piparos?), vi.20.7 ; and vishnu-as, viii.31.10. In all 
these instances the text has -os. 2. papvds, 11 (and AV. 
vii.14.3) ; pitvds, 8 ; krdtvas, 2 ; mddhvas, 12 ; vdsvas, 5 ; pipvas. 

3. The usual form in -os occurs 199 times (from 71 stems). For 
dyds, iii.54.2 (m,f.), and x.105.4,9 (m.), see p. 344, 361. For 
pikshdh, iii.19.3, BR. read piksho. 

The stems are : ahfu, 8 ; aktu, 11, and see f. ; aghdyu, 5 ; adevayu ; adhvaryu 
andfu; dnrju; ayajyu; dyu, 22; afu ; indu, 2; uru; rtdyu, 2; kdru, 7; krfd'nu. 
i.155.2; krdtu; cd'ru,1\ jantu.i; jishnu,4.; tanyatu; tapu; tr.asddasyu ; ddkshu 
ddsyu, 8 ; didhisliu ■ devabandhu ; dyu, 4 ; dru (dros and drunas) ; dhakshu 
dhrshnu ; dhenu, 2 ; ninitsu, 2 ; patdru ; parafu ; pdyu ; pipru, 3 ; piyu, 2 
purukshu, 4 ; prabhu'vasu ; prayajyu ; prayiyu ; bahu ; madhu ; mandyu, 2 
manu, 8 ; manyu, 3 ; mdiifcatu ; mrtyu, 2 ; ydjyu ; yu, 2 ; ywvdku ; ripu ; vadh- 
uyu ; vayiyu; vam; v&yu, 10 ; vifvd'yu; vishnu, 14; vidu ; farhyu; gatru, 5 
fayu, i.119.6 ; fipu; sahyu; sddhu, 2; sindhu, 13; subandhu; sushtu; sunii, 2 
suarbhdnu, 2 ; hetic. 

Transition from C : prabhds, ix.86.5. 

Transitions to the ^.-declension : cd'runas (mddasya), viii.5.14 ; 
drunas, or n., i.161.1. 

In x.46.7, (agndyo) vanarshado vdyavo na somdh, Gr. ( Wb.) seos an old genitive 
of Vdyu, assuming that the form in -av-as existed along with -o-s and -u-as. This 
is unlikely. When we have the stronger stem, we have as a rule the weaker 
ending, and vice versa ; but cf. girayas, Ab., p. 383, and rdhddrayas, p. 384 end. 
His conjecture srshtd'h, 'hastening as the winds,' Oeb. ii.516, does not satisfy. 
BR., 'inviting;' see N.p.m. 

C. Here belong : gambhuvas (rdsasyaf), i.105.3 ; sacdbhdvas. 



Gr.s.f.] Noun- Inflection in the Veda. 411 

Genitive Singular Feminine. 

A. Here belong: aktds, v.30.13 ; dhends, 3 ; vdstos, 19; vdstos- 
vastos, 2 ; sardyos; s'tndhos, 2. 

C. Here belong : 1. bhuvds/ bhuvas-pate, AV. x.5.45bis. 2. 
agruvas, 2; tanuas, 14; vadfvdas, 3; AV., tanuas, 10; -?;<Js, 4; 
vadh&as, 2. 3. pundhy&vas. 

Transitions to B: (from A) suvd'stods ; urvdrud's (BR., -■(if), 
AV. vi.14.2 : — (from C) punarbMv-ds, AV. i.27.2 (for accent, cf. 
agr&v-ai) ; prddkud's, x.4.5 ; pvagrud's, xiv. 1.44. 

Genitive Singular Neuter. 

A. Here belong : 1. mddhuas, 2. 2. mddhvas, 65 (and AV. 
vii.3.1) ; vdsvas, 38. 

3. The forms without n (36, from 8 stems) are more numerous 
than those with n (27, from 4 stems). They are: urds ; Jcshds ; 
cd'ros,2; niddhos, 13; vdsos, 8; vd'stos, 1 ; sddhds; svddds, 3 ; 
from the AV., gtiggulos (xix.38.1) ; mddhos, 6; vdsos, 3; vdsor- 
vasos, 2 ; vd'stos. 

Transitions to the w-deelension : cd'runas, 4 ; dd'nunas, 3 ; 
mddfwnas, 9; vdsunas, 11. 

Locative Singular Masculine. 

For the genetic relations of the several forms, see p. 387. 

A. 1. The stem-final is gunated and the ending is added. There 
are 8 such forms (from 1 stems) : dnavi, viii,10.5 ; trasddasyavi, 
Val. i.10 ; ddsyavi, viii.6.14; druhydvi, viii.10.5 ; pdviravi, Val. 
3.9; vishnavi, viii.3.8 ; 12.16; sUndvi. viii.57.15. 

2. The ending is dropped and the preceding diphthong receives 
a second increment. There are 30 such forms (from 19 stems) : 
akta-d; dyau, 2 ; urati, 5 ; Jcrdtau ; caraH; druhya/d; pdr$au ; 
papaii, ; pdshtigau ; ptttdkratau ; p-drati; mdnau, 2; manati,3; 
niitddrau; yddau, 3 ; prtishtigau; sindhan, 2 ; setau ; svdrau. 

The form with n is hardly authenticated in the m. In AV. 
xx. 133.3, rdjjuni seems a probable reading, MSS. rajani. 

If there is any certain instance of a L. in -d, it must be regarded 
as due to false analogy. Gr., Ueb. ii.S12, sees such a form in 
ix.97.37, rtdl (for rta'd) matmd'm, 'at the time of prayer.' In 
AV. xx.131.12, the MSS. have vanishthd', ed. -ati. 

C There is no example. 

Locative Singular Feminine. 

A. The Rik has only sindhau, viii.20.25 ; the AV., sdrau, 
v.25.1 ; stZ'tau, i. 11.1. 

C. 1. The VS. has bhruvi, xix.91. 2. Of the organic form 
there are 4 examples : cam&i; tan-di, 3 ; and with metrical length- 
ening, 3 : tanM, p. -i, ii.16.2: x.65.7 ; 66.9. The metre makes 
against the lengthening in iv.6.6, tanM (Prat, viii.7), p. 4. Cf. 
stems in r (ar), L.s. 



412 G. R. Lanman, \u and M-stems. 

The ending is sometimes dropped (or absorbed by the thematic 
vowel?). As a L., camil' occurs 6 times; but in x.24.1, it may 
be instr. In x.183.2, svd'ydm tanti,' r'tvie (L.s.n.) nd dhamdndm, 
tanH' is L.s. : ' praying (for a son) in thy body at the time of the 
rt&: The -d of the L. is pragrhya, Prat. i.28. Cf. p. 389, B3. See 
note to Ath. Prat, i.74: tawQl f, iv.25.5 ; mdyd', xviii.4.4. 

Transitions to B. The Rik has only pvapned'm. The AV. has 
rdjjvdm, vi.121.2. We ought probably to read tanvd'm : m 
i.18.3, ed. -vd"m ; in xix.61.1, tand's tanvd'm me bhaved antdh, 
MSS. -yd', ed. -vd"m; and in 37.2, text -vdm; and, possibly, tanud'm 
in 55.3, text -vdm. Cf. note to Ath. Prat, iii.60. Similarly Roth 
has proposed for Rik v.58.7, sud'm, text swam. 

Locative Singular Neuter. 

A. 1. Here belongs sd'nav-i. It occurs 9 times. 

1. (a) Furthermore, the combination sd'no dvye occurs 7 times, 
and sd'no avydye once. In the last instance the -o occupies the 
ninth place of a jagati ; in the other 7, the ninth of a trishtubh. 
The pada has -au in all 8 instances, and the Prat., ii.35, implies 
that the natural sarhhitd-form would be -dv. The metre shows 
this view to be wrong, since the -o has the value of a short syllable 
(dv). I think that the -i or -y of the ending has been simply 
dropped before a vowel, in order to avoid a cacophony (e. g., 
sd'ndv-y avydye) ; and that this has happened without the usual 
compensatory lengthening, in order to give the requisite cadence 
-w-s_- or - - - -. 

2. Here belong 28 forms (from 8 stems) : arajjati, ; ura-d, 6 ; 
ghr'shau; prthad; mddhau, 4 ; vdsau, 3 ; vidati, 2 ; sd'nau, 10 : 
AV. xi.5.12; mddhau, AV. ix.l. 16,17. 

Transition from C : (pdrman) udbhaii, VS. xv.l. 

Transitions to the w-declension : d'yuni, 3 ; sd'nuni, i. 155.1 
(Gt., --Am); dd'runi, AV. vi.121.2. Jacobi, Ind. Stud, xiv.144, 
cites amuni=amushmin; cf. vipve=vipvasmin, iv.16.19, and Gr. 
1299. 

Bollensen, Orient und Occ. ii.481, proposes for vii.39.3 urd'n- 
tdrikshe, i. e. urd' an-, text urati,. Cf. masc. 

Yocative Singular Masculine. 

A. The stem-final is gunated : thus, adhvaryo. The form 
occurs 508 times (from 58 stems). Examples are : indo, 144 ; 
dhrshno, 19; vaso, 62; vdyo, 43; vishno, 17; patakrato, 47; 
sukraio, 22 ; stino, 36 ; further, adhrigo, 3 ; sishno. BR. read 
piksho, iii.19.3 ; see G.s.m. 

In AV. iv.7.3, dushtano is possibly a transition-form (for 
dushtawti), since the Vedic stem is properly dushtanW. 

The form sahasracaksho (V.s.m.f., for shus), AV. iv.20.5 : 
xix.35.3, may be regarded as analogous to bhadra-$oce, etc., p. 390 
top. We might assume a shorter stem in -u equivalent to that 
in -ns; but it is hardly necessary or warrantable here. 

C. There is no example. 



V.s.f.] Noun- Inflection in the Veda, 413 

Vocative Singular Feminine. 

A. Here belong: adrisdno; durha.no; prthushto; sindho, 3; 
suhdho / svabhdno. For sahsracaksho (os/iadhe), AV. iv.20.5, 
see m. 

C. Here belong: jtihu, AV. xviii.4.6 ; vadhu, xiv.1.58; 2.7; 
bdbhru, vi.139.3. 

Vocative Singular Neuter. 

A. The Atharvan has guggidu, xix.38.2. 

Nominative, Accusative, and Vocative Dual Masculine. 

A. The genesis of the form is uncertain (p. 390). It ends in -ti: 
as,adhvart/H'. It occurs 196 times (from 69 stems). Examples are: 
indr&mdyd ', 22; \ndrdo%shni%, 13; bdhti', 20; vdjinivasti, 21; 
vr'shanvasit, 18 ; further, dsmrta-dhru / rta-psu ; ghrtd-smi 
(sdnu), 2 ; vanar-gil' '/ vrdhasnil'. 

Transitions from C : puru-bhO! ; pambhii'; pambhd; paribhic', 
AV. iv.25.1. In this passage, the restoration of the older and 
longer grammatical form (the true C-form) is strongly favored 
by the metre : thus, yau vipvasya paribhtivd babhUv&thuh. 

The form appears in both texts with shortened final : jigatwa! 
(mitrd'vdrund), vii.65.1 ; suhdntii, (cumurirh dhunirh ca), 19.4. 
Cf. p. 390 ad fin., 391, B. 

C. Here belong : manojuvd; mayobhi'tvd, 7 ; pambhiivd, 2, and 
vi.60.14; sacdbhuvd, 6. Almost all are at the end of a pdda. 
Here I would put ghrta-snuvd, iii.6.6. See p. 402 top. 

Transitions from A: madhilyuvd, 2 ; papvd', x. 1 06.3 (p. 403 med.). 

Nominative, Accusative, and Vocative Dual Feminine. 

A. Here. belong: jigatnil' ; dhenif, 4, and see p. 406 ; sdbandhu, 
2 ; samdndbandhil ; hdnil, 4. 

C. Here belong : 1. srdvd; apijdvd; punarbhuvd; pambMivd; 
sacdbh/dvd ; vipvdpdmbhuvd, i. 160.4: vi.70.6 : i. 160.1 (-uvd r-, a 
graphic peculiarity; cf. p. 342 and Prat. ii. II): — Mh/mm, Qat. 
Br. xiv.9.4 5 . 2. canvdd, / tawdd, 6. Form as well as sense 
forbid our taking tatM' (r'tvie) as dual, x.183.2. See L.s.f. 

Nominative and Accusative Dual Neuter. 

A. The Rik has only urvf, 3. As these cases are " weak" in 
the neuter, the accent of an oxytone stem (urd) is shifted to the 
vocalic ending (urv-t ', not urv-f) ; cf. papv-ds, pitv-ds, etc. The 
VS. has jd'mi-n-i, xx.8. 

Instrumental, Dative, and Ablative Dual Masculine. 

A. Instrumental : andmayitm'ibhydm , nrbdlvdbhydm ; b&Mi- 
bhydm, 6. Dative : bdhtibhydm. Ablatives : lirubhydm ; bdhu- 
bhydm. 

Instrumental, Dative, and Ablative Dual Feminine. 

A. C. The Rik has no example. VS. xxv.l, bhrti-bhyd'm. 
Transition from A to C : hdntfbhydm, p. it, TS. vii.3.16'. 
vol. x. 56 



414 0. R. Lanman, \u and w-stems. 

Instrumental, Dative, and Ablative Dual Neuter. 

A. C. The Rik has no example. The AV. has jd'nubhydm, 
ix.8.21: x.2.3. 

Genitive and Locative Dual Masculine. 

A. Genitives: Uruds; bdhuds, v. 16.2. Locatives: bdhuds, 21 
times; tirvds, AV. xix.60.2; bdhvds, vii.56.6 : xix.60.1 (ed. -vds!). 

Genitive and Locative Dual Feminine. 

A. Locatives: hdnuos, i. 52.6 : AV. x.2.8; -vos, x.2.7. 

C. Locatives: 1. bhruvds, iv.38.7 : Cat. Br. xii.9.1 6 . 2. 

caiwdos, 13 times, and ix.69.5 ; 108.10; oamvds, ix.96.21 (hut see 
p. 408 top). 

Genitive and Locative Dual Neuter. 

A. C. The Rik has no example. AV. s..2.2,jd'nu-n-os. 

Nominative and Vocative Plural Masculine. 

A. 1. Here belongs only mddhu-as, ix.89.3. 2. As a dis- 

syllable, mddhv-as occurs thrice. I would add Vibhv-as — taking 
it as vocative of the stem Vibhib'—, at the beginning of iv.34.9«? 
(Gr., N.p. of stem vibhti) and vii.48.lt?, and Vibhv-as, iv.36.3. 
Cf. VibhU'-bhis, vii.48.2. The stem VibhiZ,' is perhaps differen- 
tiated as a proper name from vibhtf,'; cf. p. 368 top. 

3. The stem-final is gunated: as, aktdv-as. The form occurs 737 
times (from 161 stems)! Examples are (N. and V.) : adhvarydvas, 
27; dydvas, 27; dpdvas, 30; indavas, 67; rbhdvas, 57; kdrdvas, 
18 ; vdsavas, 46 ; smdhavas, 34 ; sudd'navas, 42 ; further, d-psavas, 
and compounds of 2-psu, -dyu; dgraoas ; mitd-jnavas, 2; rnitd- 
dravas,2; hari-dravas (p. 401). Here belong probably : vdydvas 
(sdmds), x.46.7 (cf. p. 410 end); and vdydva ind-, p. -we, vii.92.4. 
See BR., Zvdy-d. 

Transitions from C : dhi-jdvas; vibhdvas. 

In i.64.3 and viii.22.11, we have from the stem ddhri-gu (-gu— 
gam), ddhri-gdvas. The d is perhaps due to the false analogy of 
gd'vas, ' cows.' Cf. D.s.m., p. 409. 

C. Here belong: vibhuas; subhuns,%; surdpi'ias; and, d-duvas; 
dbhuvas, i.64.0 : ix.65.27; u-h-dvas ; j-dvas, 2; nabhojuvas; pari- 
bhtivas; mancg'tivas, 3 ; mayobh'uvas, 6 ; raghu-drdvas, 3 (see p. 
402) ; vayoj-dvas; pambh&vas; sadyoj-dvas; sudbhdvas, 4 ; mayo- 
bhtivas, AV. vii.60.2 ; subh-dvas, iv.8.7. 

Transitions from A: dprdyuvas; madhydydvas; mitrdydvas; 
pramaydvas. 

Nominative and Vocative Plural Feminine. 

A. 2. Here belong : mddhv-as, Val. 2.4 ; patakratv-as. 3. Here 
belong 89 forms (from 15 stems). They are: anaprdvas ; dbhi- 
ravas; arendvas; tshavas, 3 ; tridhd'tavas ; dhendvas, 49 ; pdr$- 
avas,2; vasHydvas, 3 ; pdravas ; sanishydvas, 2; sdmanyavas; 
smdhavas, 21 ; suketdvas; svdsetavas; hdnavas. 



N.V.p.f.] Noun- Inflection in the Veda. 415 

C. Here belong: 1. ghrta-pti-as ; navas'uas,i\.S4.5; prastias,^; 
oibhuas ; sublv&as: — d&vas; dhijuvas / punarbh/dvas ; bhuvas • 
mayobhtioas, 3; mitra-kruvas (p. 402, 1.3.). 2. (a) jatuas, 

AV. ix.2.22 ; prddkuas, AV. i.27.1 ; agruvas, RV., 6 ; (b) camtias; 
ljuhuas, 4 ; ijulvuas ; tantias, 2, and x.108.6 ; tanvds, x.51.4 (p. 
408 top) ; tanuas, AV., 4, and xiii.3.16 (-a"s, Ath. Prat, iii.65, note, 
end). 3. didhishuas y pdrayishw&as ; mumukshiias; sUdayit- 
nuas; prajanuas, AV. ix.4.6 : — anhoyuvas; apasydvas, 4 times, 
and ix.2.7 ; avasydoas ; dyuvas • udanyiivas ; panasydvas ; 
prtandydvas; prpandyuvas; bibhatsiivas; makhasydvas; man- 
draydvas; mahiyHvas, 2 ; sandy "dvas ; caranydvas (giras), AV. 
xx.48.1. 

In iv.41.8, yuvayd's is certainly to be taken with dhiyo, N., and 
must therefore stand for -yd' -as — an instance of vowel-absorption 
common with the 2-stems (p. 366). Cf. A.p.m., jU's. In x.70.5, 
we have rathaydr (dvd'ras) ; and in vii.2.5, rathaydr {dtiras). Gr. 
reads in both places -yd's {-yd'as); but BR., vi.257, consider it a 
syntactical peculiarity. 

Nominative and Accusative Plural Neuter. 

A. 1. The samhitd has -it; the pada, -ib. Here belong 28 forms 
(from 4 stems) : urDJ ; purtt', 24 : as, iii.51.5, purOl vdsilni prthivi' 
bibharti; vdsti, 2 ; vidd' (utd, Prat, ii.37), i.39.2. 

2. Both texts have -■&. Here belong 48 forms (from 12 stems) : 
urh, 2; rju, iv.1.17 : vii.60.2 ; cdlru (dnnd), i.61.7; tridhd'tu ; 
purd, 11 times (at end of pdda), and vi.44.14; bahti; mddhu, 2, 
and iv.43.5; vdsu, 19 times (12 at end of pdda); vtdu, 3; sd'nu, 
vi.61.2 ; viii.85.2 (trih saptd) ; sudd'tu; suhdntu, vii.30.2 (but see 
I.s.n.). In vii.60.3, (dhd'rndiii) yuvd'kuh s-, the visarga seems out 
of place. See p. 406 top. I believe the AV. has only one such 
form, purd', xix.49.4. 

Transition from C : (bheshajd') mayobh'd, ii.33.13£> end. 

Transitions to the w-declension. Here belong 127 forms (from 
14 stems) : cd'rHni; tridhd'tUni, 2 ; dd'niXni; dariJLrii; devayd'ni/ 
purd'ni, 39; prthti'ni, 2; bah/dlni ; mddhdni, 9; yuvayd'ni ; 
vdsilni, 66 ; vd'sMni; pmdprtlni; sd'nUni; from the AV., arydJni, 
xi.7.10; ald'bilni, xx. 134.1 ; dprtini, v.19.13. 

Accusative Plural Masculine. 

A. 1. Here belong: papv-ds, 4; kr'tv-as, 2 (BR., s.v.). 

2. The usual pada-iorvo. in -tin occurs 189 times (from 43 stems). 
(a) If it is followed by a vowel in the interior of a pdda, the sam- 
hitd shows the phonetic representative, -Unr, of the original case- 
form -Uhs. See p. 346, 394, and Prat, iv.29. This occurs before a- 
31 times : as, rtti'nr dnu, i.49.3 (but kdrti'n \ dhnd, iv.16.3) ; before 
d-, viii.57.16 ; before iva, vi.46. 14 ; 57.6: x.68.2 ; before i-, i.45.1 : 
ix.97.17; before «-, v. 31.1 3 ; 42.15: x.83.6 ; before r-, x. 2.1 ; before 
o-, vii.5.6; before e- (at the end of the pdda, Prat, iv.30), vi.18.3a/ 
and even before y- (both cases are covered by the Prat., iv.29), 
i.63.4: v.42.15. Total, 45. (b) It occurs in the interior of a 



416 G. R. Lanman, [u and ttstems. 

pdda : in 36 instances, before consonants, unchanged ; further- 
more, once as -tinp, before ca, in i.100.18 ; as -'(In, before ca, i. 72.6 ^ 
as -fin, before j-, v.14.4; as -tin (ch-), before p-, i.100.18 : Val. 7.3. 
(c) It occurs at the end of a pdda: as -An, before o, i. 174.6; as 
-tin, before j-, vi.44.17 : x.180.1 ; and unchanged in the remaining 
1 00 instances. For the Atharvan usage and statistics, see note to 
Prat, ii.29. 

The stems are : aktu, 9 ; akrata ; addfu ; adevayu, 2 ; abM'fu ; ayajyu, 2 ; dpu, 
3; dsannishu; indu; rtu, 8; rbhu; kdru, 2 ; kratu; tantu; dasyu, 21 ; dd'nu; 
diu, 25; durhrndyu ; nabhanu; parafu; pafu, 5; prtandyu, 3; prtanyu, 4; pra- 
yojyu ; bdndhm ; babkru ; bdhu, 2 ; mitreru ; y&jyu ; yuvanyu ; vagnu ; vanu ; 
vard'hu; v&su, 3; vdyu; venu; vetasu; fdtru, 43; fimyu; sindhu, 23; sudd'nu ; 
sO/nu ; suabhtyu. 

Transition from C : mayobhU'n, i.84.16=AV. xviii.1.6. 

C. Here belong : dndbhuvas; ntayobhUvas. 

In ii.14.3, indram sdmair Srnuta jil'r nd vdstraih, we have 
probably an A. p.m., for ju-as. Cf. yuvayti's, p. 415, and see BR. 
iii.128. 

Accusative Plural Feminine. 

A. 1. Here belongs : mddhv-as, i.181.6 : iii.31.16. 

2. Here belong : uhiJLs; ghrtd-swds, ii.27.1, cf. C ; dhewd's, 4 ; 
vdreniakratds, in a khila to x.9 ; ishds, AV. xi.6.9. 

In x.35.2, (mdtf'n) svndh/dn, we have a masculine form with the 
signification of a feminine. In x. 11 1.9, etd's, f., refers to sindh&n 
jagrasdnd'n. 

C. Here belong: 1. prasAas, 3; vibhUas ; dbhuvas ; dhtivas ; 
ghrta-muvas, i. 16.2 ; duvas, 2 (see also duvas, Gr. 617) ; bhuvas; 
mayobhuvas; sandjuvas. 2. tanuas, 19 times, and v.15.3 (Say., 
svds tanus); nabhanuas; vadhuas ; (from the AV., tanuas, 4 ; 
-vds, 1; badhuas, viii.6.14;) agr&vas. 3. abhidipsuas; dravit- 
whas; vipanyuvas; pundhytivas. 

It seems necessary to pronounce tanvds as a dissyllable in 
i.162.20 and x.51.2 ; see p. 408. Weber, Ind. Stud, xiii.58, con- 
jectures with some doubt vipvd (as A.p.f. of a stem vipu', ' nest') 
for vipvo, ii.38.8. Although I do not know of any other occur- 
rence of such a stem, it seems to me a perfectly warrantable 
formation from the root vip. 

For rathayu's (duras), vii.2.5, text -us, see p. 415. 

Instrumental Plural Masculine. 

A. The ending is added directly to the stem : as, anpu-bhis. 
The form occurs 210 times (from 50 stems). Moreover, Vibhu'-bhis 
occurs vii.48.2. It might be regarded as a transition-form ; but 
see p. 414 med. The long u- (Prat, vii.2) of makshu'-bhis, p. &, viii. 
26.6, is due to the metre. BR. take the word as an adv. instr. 
Say. says: pighragamanair apvaih. For the accent of snu-bhis, 
cf. s[d']nu-bhis, vii.88.3, and see p.' 408 med. The AV. has vanar- 
gubhis, iv.36.7. 

The stems are: ancA, 3; aktu, 11; adhvaryu, 8; apasyu ; aprdyu; abhidyu, 3: 
abhi'git; arajju; arena; avasyu; dyu,i; drujatnu; dfi,,\2; indu, 11 ; uru, ; rtdyu, 



I.p.m.] Noun- Inflection in the Veda. 417 

3; rtu, 15; rbhu, 6 ; kratu, 11 ; gdtu; jantu,2; jigyu; tdntu; tarn; tr'tsu, ; ivdyu ; 
dyu, 19; devayu; namasyu; pdyu, 12; prushitapsu, 2; ftaftii, 4; bhdnu, 7; bhiru; 
bhr'gu ; makshuyu ; manyu, ; mitajfhc, 2 ; vacasya ; v&su, 24 ; vdyu, 4 ; vipanyu, 
2 ; sanishyu, ; saranyu, 2 ; sddhu, 2 ; sindhu, 5 ; sudd'nu ; svtyantu ; sunu ; s»w, 5. 

C. Here belongs dbhii'bhis, 2. For makshU'bhis, see above. 
Instrumental Plubal Feminine. 

A. Here belong : kkadhenubhis ; tridhd' tubhis ; dhenitbhis, 5. 
C. Here belong: sva-pU'-bhis ; Xjuhil'bhis ; 2juhii'bhis, 2 ; tan- 
d'bhis, 9 ; AV., 4. 

Instrumental Plural Neuter. 

A. Here belong : arenubhis, vi.62.6 ; baMibhis ; vdsubhis, 3 ; 
sumdntubhis (stotrais, Say.), i.129.7; jard'yubhis, AV. i.27.1 ; , 
bahUbhis, vii.26.8. 

Dative Plural Masculine. 

A. Here belong 18 forms (from 11 stems) : apatrubhyas; rbhu- 
bhyas; gungubhyas; tr'tsubhyas, 2 ; ddpdbhipubhyas; ddsyubhyas, 
iv.38.1 ; papubhyas; pilrubhyas, 2; bahiibhyas, 3; vdmbhyas, 2 ; 
sindhubhyas, 3 ; from the AV., papicbhyas, 2 ; bahiibhyas, 2. 

Dative Plural Feminine. 

A. The Rik has no example. The AV. has ishubhyas, iii.27.1-6 ; 
dheniibhyas, vi.59.1. 

C. Here belongs tanti'bhyas, x.158.4 (Gr., -bhias) ; AV., i.13.2 ; 
26.4. 

Dative Plural Neuter. 

A.C. There is no example. 

Ablative Plural Masculine. 

A. Here belong : aktubhyas; jatrubhyas, viii.1.12 (see Gr. 548); 
ddsyubhyas, x.48.2 ; bahiibhyas; bhr'gubhyas ; from the AV., 
jighatsubhyas, viii.2.20 ; ddsyubhyas, ii. 14.5 ; bhr'gubhyas?, xix. 
39.5; mrtyubhyas, xiii.4.46 ; sdbandhubhyas, viii.2.26. 

Ablative Plural Feminine. 
A. Here belong : dhdmtbhyas; sindhubhyas, 3. 

Ablative Plural Neuter. 
A.C. There is no example. 

Genitive Plural Masculine. 

A. So far as I know, the Veda shows no certain exception to 
the rule that oxytone stems belonging to A shift the accent to the 
ending in the G.p. : thus, babhrilnd'tn, m. On the other hand, 
the thematic final of words belonging to C retains the accent : 
thus, babhr&'ndm, f. The accent of the i and «-stems is quite 
analogous. From the A-stem kavi, m., we have kaotnd'm; but 
from the C-stem krimt', f. (m., krimi), we have krimi'ndm, AV. 
v.23.13. See pp. 397-9. 



418 G. R. Lawman, [u and w-stems. 

Oxytones. The Rik has 24 forms (from 12 stems). 1. rtdnd'm; 
rbhilnd'm, 2 ; rsh'dnd'm ; cardnd'm ; devayund'm ; pitdnd'm; 
prdpdnd'm ; babhrund' in ; bah'dnd'm, 3 ; ydttin&'m. The AV. 
has: rMnd'm, xv.6.6 : xvi.8.17; rbhibnd'm, ix.1.13; cardnd'm, 
xviii.4.53 ; devaydnd'm, viii.9.13 ; papilnd'm, 24 times : as, ii.28.3 ; 
ydMnd'm, v.29.8,9. The VS. has stdydnd'm, xvi.21. For aghd- 
yd'ndm, AV i.20.2, a possible exception to the rule, see G.p.f., 0. 

2. The following words occur at the end of a catalectic pdda of 
7 syllables, where Gr. reads -dam : rbhdndlm ; rshUnd'm, 3 ; 
pitdnd'm; purdnd'm, i.5.2 ; 36.1: v.74.7: vi.45.29 ; ripHnd'm; 
further, rshilnd'm, at the end of the trishtubh-pdda, i. 12 7. 10/". 
3. In no case is resolution certainly necessary. 

Barytones. There are 23 forms (from 11 stems). 1. abhikra- 
tUndm; abhi'ptlndm; ahydrshdndm; tr'tsilndm, 2 ; bhr'g-dndm; 
tndwdndm; vdsUndm, 5 ; $dtrtindm, 2 ; sindkdndm, 6 ; svdrdndm. 
The AV. has: ddsydndm, iii. 10.12 : viii.8.5,7 : x.6.20; piydrHndm, 
xi.2.21 ; vds&ndm, ix.1.4 ; g,dtrdndm, iii.19.2 : v.20.4 : vi.65.2 ; 
sabandhUndm, xv.8.3. 

2. At the end of a pdda of 7 syllables occurs krdtdndm, Gr., 
-aam. 3. Resolution seems necessary in sindh/dnaam, ix.15.5. 

C. Here belong: ydtujti'ndm,'\\A.5: x. 116.5 ; joguvdm. 

Genitive Plural Feminine. 

A. Oxytones. 1. The Rik has no example. The AV. has 
dhenilnd'm, iv.27.3. 2. No example. 3. Resolution is neces- 
sary in dhenHndam, RV, 3. 

Barytones. 1. Here belongs slndhdndm, 5. 3. Resolution 
is necessary in smdhdnaam, i. 46.8.9. 

C. Here belong : 1. p-drvasti 'ndm. 2. tanH'ndm, 12 ; tan-d'- 
ndm (Gr., -aam), v.67.5: vi.48.2 ; vadM'ndm (Gr., -aam), v'm. 
19.36 ; tanH'naam, ii.23.8. 3. babhrd'ndm, x.97.1 ; Mbhat- 

s'd'ndm, x.124.9. If we judge aghdyd.'ndm, AV. i.20.2, by its 
accent, it is a feminine from the stem aghdyd' (m., aghdyd'), and 
may mean malicious beings of that sex. If it be taken as a mas- 
culine (for aghdydnd'm), it is an exception, and the only one, to 
the rule in regard to the shift of accent. 

Genitive Plural Neuter. 

A. Oxytone. For purilnd'm, i.5.2, see masc. 

Barytones. 1. Here belong : mddhdndm, 6 ; yd'pdndm; vdsrd- 
ndm, 25, and AV. iv.26.1,2 : vii.79.3. 2. At the end of a cata- 
lectic pdda of 7 syllables, where Gr. reads -aam, stand : mddhil- 
ndm; vdstindm, i.127.7 : viii.31.14: ix.58.2: Val. 3.5 ; and at the 
end of a pdda of 1 1 syllables (where Gr. reads -aam, 12), vdsdndm, 
vii.16.2 ; 32.5. 3. Resolution is necessary in vdsdnaam, i.7.9; 
128.5: ix.108.13: x.50.7 ; 74.1. 

Locative Plural Masculine. 

A. Here belong 17 forms (from 10 stems) : aktushu ; dnushu; 
dpushu; druhydshu; ptirushu, 2 ; bdkdshu, 3 ; yddushu; vipvd- 



L.p.m.] Noun- Inflection in the Veda. 419 

bhdnushu, 2 ; pdtrushu ; sindhushu, 4. The A V. has : paptishu, 
6 ; vibandhushu; the VS., snushu, xvii.14. 
G. Here belong : dblvd'shu; purubhu-'shu. 

Locative Plural Feminine. 

A. Here belong : vdstushu; sindhushu, viii.24.27. 

Transition from C : (yikshu) dyd'shu, i.58.3. We should expect 
dyd'shu; cf. dyuvas, ii.5.5. The & is perhaps metrically shortened, 
as eleventh syllable in a pdda of 12. 

C. Here belong : 1. prasii'shu, twice. 2. camU'shu, 13 ; 

tamO/shu, 20, and AV. xix.20.3. 3. babhrv'shu. 

Locative Plural Neuter. 

A. Here belong : urwshu ; pmdprushu ; sd'nushu, 2, and AV. 
x.4.14; vd'stushu, xii.5.49. 



STEMS IN R OR AR. 

In the word's of which this section treats, the distinctions 
between "strong" and "weak" stems come for the first time 
prominently and unequivocally into notice. The stem ddtd'r, 
' giver,' makes its A.s., ddtd'r-am, from a fuller (" strong") stem, 
ddtd'r; but its L.s., ddtd'r-i, is formed by adding the ending -i to 
the simple (" weak") stem, ddtd'r. The cases which exhibit this 
fuller form of the stem are called, in general, "strong," and are, 
for the masculine and feminine, the N. and A.s., the N.A.V.d., 
and the N.V.p. ; the others are called " weak." This strengthen- 
ing of the stem does not extend to all words included here, and 
they may therefore be divided into two classes (A and B), accord- 
ing as they have dr or dr in the strong cases : thus, md'tar, 
' measurer' (A.s., md'tdram), may be called an A-stem ; and mdtdr, 
' mother' (A.s., mdtd'ram), a B-stem. 

A. This category embraces 150 or more stems, formed by the 
derivative suffix tar, which are used as nomina agentis, or as 
participles. In general, the former are oxytone, and the latter 
accent the radical syllable. Lindner, p. 72, is of the opinion that 
the difference was originally a purely formal one, and that this 
formal distinction was afterwards extended and utilized as the 
expression of a functional difference. Here also we may put the 
stem star or tdr, 'a star,' although it is properly a radical. It 
occurs in the N., td'ras, and I., str'bhis. 

B. This class is a small, but important one. It includes a 
number of words of relationship — substantives and their adjective 
compounds. The masculine stems are : the compounds of -mdtar, 
' mother,' with ddri-, ih&ha-, g6-, pr ' $ni-, saptd-, sindhu-; trimdtdr, 
dvimdtdr, sammdtdr, sumdtdr • jd 'mdtar, vijdmdtar ; deodr / 
(ndplar, ' descendant,' supplies its strong cases from a different 
stem, ndpdt; but cf. naptdrem,Ya,<}na,lxxi.23;) pitdr, ddkshopitar, 
mdtdrd-pitdr • bhrdtdr [saptdsvasar]. The feminine stems are : 



420 C. B. Lanman, [Stems in 

duhitdr ; ndndndar ; abhrdtdr ; mdtdr, saptdmdtar, sammdtdr, 
s'lndhumdtar [svdsar, saptdsvasar]. For ndr, see N.p.m. 

C. There are only seven stems included in this section which 
are not formed by the derivative suffix tar. They are : itshdr; 
devdr; ndndndar; ndr, stiarnar; svdsar, saptdsvasar. Of these, 
ushdr, ndr, s&arnar, and svdsar have peculiar declensional forms : 
usr-ds (G., A.), -d'm (L.) ; ndr-e, -as (G.s.), nar-d'm; s'darnar-e ; 
svdsr-drn (G.p.). 

Among the weak cases, moreover, there are certain changes of stem which 
require a descriptive notice. The stems treated in this section — like those in i 
and t, and u and 6, — have a final that lies on the border-land between vowel and 
consonant. If we consider them chiefly in the light of the related languages, we 
must entitle the section "stems in tar and ar ;" but this declension exhibits 
peculiarities so specifically Indian, that I have followed the usual method, and 
given it a place with the declension of stems whose suffixes end in vowels (p. 327). 

From a comparative point of view, we say that the stem of ddt&r-i is ddtar, and 
that in ddtr'-bhis we have a peculiar weakening thereof ; but if we compare 

agnibhydm -ibhis -ibhyas -ishu, 

with ddtr'bhydm -tr'bhis -ir'bhyas -tr'shu, 

and Ague agne-s *agndy-i agndy-e 

sd'no sd'no-s sd'nav-i sd'nav-e, 

with dd'tar *ddtar-s ddtar-i *ddtdr-e, 

we see that, as far as the Sanskrit is concerned, agni and sd'nu stand on the same 
phonetic level — not with ddtar, but with ddtr'; and that, taking ddtr' as stem, we 
have in ddtar-i a peculiar strengthening (guna) thereof. Likewise the lengthen- 
ings in the A. and G.p. are quite analogous (agni'n, -ind'm — ddtr'n, -tfnd'm). 
Further, the analogy of urmi-d' and -agni-os would lead us to assume, for the 
period of separate Indian development, the forms ddtr-d' and ddtr-6s (not ddtdr-d, 
ddtar-os). In the Rik, the G. and L.d. forms are trisyllabic in every instance but 
one, and may be pronounced pi-tr-6s (Gr., pitaros), etc. 

In Sanskrit it is tho rule that the & of thematic tar falls out in the D.s. (e. g., 
ddtre, for ddtare), and remains in the L.s. In the Rik, the metre does not give 
evidence of a single form like ddtare; but in x.85.46, the text has nandndari 
where the metre demands -dri (A V., -dm), and, on the other hand, usri, v.53.14, 
where the metre demands ushari. 

In the weak cases, accordingly, the stem appears in four forms : tr, tf, tar, and 
[<(«)r, i. e.] tr. The fourth form, ddMp,')re, becomes by syncopation ddtre. and the 
resultant is thus similar to that of the first form ddtrd' (from ddir-d'). 

There are a few instances in which the distinctions between A 
and B are ignored ; but they are too few to be classed as transi- 
tion-forms. The most important irregularities are those of the 
stem svdsar, ' sister,' and its compounds. Although a word of 
relationship, it has no B-forms whatever, but makes svdsdram-, 
svdsdrd, svdsdras. Some recognize in the peculiar declension of svd- 
sar traces of its original character as nornen agentis (*sva-siX-tar). 

The instances in which A-stems have B-forms are only sporadic. 
We have in a pdda of 1 syllables, manot&'rd rayind'm, viii.8. \2b 
=i.46.26. In iv.35.5, fdcyd hdri dhdnuttirdv atashta, it is likely 
that the d, is due to the metre. The word vasudhd'ttiras, AV. v.27.6, 
is discussed in the note to the Ath. Pr. iv.45. Some authorities 
regarded it as a plural. This interpretation, which is favored by 
the sense, is not so very objectionable on account of the accent ; 
for we may take the word as a compound of the participle dhd'tar. 
This occurs in Rik viii.'7.35 ; and the compound sdmdhdtar, viii. 1.12, 



r or ar.] Noun- Inflection in the Veda. 421 

implies the barytone dhd'tar. Otherwise, it must pass for a 
N.s.m. of the comparative of vasu-dhd 1 . 

Except in the G.p., no forms of this declension show a n be- 
tween stem and ending. For str-n-as, see N.p. 

The following synopses exhibit the terminations of the in- 
flectional forms : 

Neuter. 

Singular: N.A., -tar?, -tr?, -tur, -tari ; G., -tur ; L., -tart 
Plural: G., -trdm ?. 

Masculine and Feminine. 

Singular: ~N.,-d,-dn,-dr?; A., -dram, -dram; I., -rd ; D., -re 
(-are) ; Ab., -ur; G, -ur (-ras, -aras) ; L., -art, -art (-ri, -rdm) ; V., 
-ar. 

Dual : N.A.V., -drd, -drd, -drau, -drau; I)., -rbhydm; G.L., -ros, 
-r-os. 

Plural : N.V., -dras, -dras; A.m., -rn, -fnp, -fnr, -fnh (-aras?) ; 
A.f., -fs (-ras); I., -rbhis ; D.Ab., -rbhyas; G., -fndrn, -rnaarn, 
-rndm, -rnaarn (-ardm, -araam, -rdm), -rn ?; L., -rshu. 

The forms in parentheses pertain only to the stems ushdr, ndr, 
s'uarnar, and svdsar. 

The following peculiarities are seen in the accent of the forms 
of oxytone stems. It is shifted to the ending in the I.s. and 
G.L.d. (pilrd', mdtros — for pi(r-d', mdtr-os — p. 420), and in the 
G.p. (pitfnd'm). It is also shifted to the ending when the final 
syllable of the stem loses its character as a distinct syllable by 
syncopation ; as, ddtr6, usri (for ushdr -i). Compare the accent 
of the stems in i and w, pp. 375 and 405. 

Contrary to the general rule that monosyllabic stems shift the accent to the 
ending in the oblique cases (e. g., bhi', bhd'), we have : ndr-e, -as, -i ; nr'-bhis, 
-bhyas, -shu (but nar-d'm); str'bhis; and in like manner, gav-d, -e, 4, -dm; go-bhis, 
-bhyas, -shu. The stem nau follows the rule (ndv-ds, -i; nau-b 1 -' ■•■' 



This declension pertains almost exclusively to masculines and 
feminines ; and this is to be expected, since the words in tar are 
the names or epithets of persons or personified objects (Lindner, 
p. 72). The explicit paradigms of the grammarians (cf. bahusva- 
sr'ni) have little or no application to the Veda. Least of all 
should elaborate arguments be based upon facts with regard to 
the forms of the neuter. 

NEUTER STEMS IN TAR. 

The Kigveda, however, appears to have a few forms (perhaps 
1 7, at most), from neuter stems in tar. Since they are so few and 
sporadic, it seems advisable to give them all together here. The 
stems are : dhartdr, ' prop, support ;' dhmdtdr, ' smeltery, smithy ;' 
sthdtdr, the opposite oijdgat; and vidhartdr, used as an infinitive 
' to hold (out),' ' to mete out.' ' 

vol. x. 57 



422 G. R. Lanman, [Stems in r. 

Nominative and Accusative Singular Neuter. 

In the first place we should notice the fact alluded to on p. 343 
(cf. p. 37 7 end, 412 end, and 390, and Gr., Wb. 1603), that in cases of 
extremely infrequent occurrence there appears to be a correspond- 
ing uncertainty as regards the form. It would seem that linguistic 
usage was to that extent still unsettled and indeterminate. For 
the N. A.s., the Veda shows at least two, and perhaps four attempts 
to make case-forms. 

1. The termination that we should expect d priori is -tdr. The nearest 
approach to this is in vi.49.6, sdtyagrutah kavayo ydsya gtrbhir jdgata (p. -tali) 
sthdtar jdgad d' krntidhvam, ' At whose call ye. truth-hearing wise ones, ([are] 
coming, i. e.) come, (to him) grant ye that which remaineth steadfast and that 
which moveth,' i. e., make him master of all things. The word sthdtar, however, 
has no accent. Gr. reads -ur, and BR. pronounce the passage corrupt. 

2. The form of the paradigms ends in -tr' ; as, kartr'. BR. suggest the 
reading sthdtr' for the passage pafU'n ca sthdtr'n cardtham ca pdhi, i.T2.6, and, as 
I think, with good reason ; since all masculine forms (except this) come from the 
barytone stem sthd'tar. Cf. Miiller's note to Rik Pr. iv.32. 

3. As a phonetic representative of sthdtar or sthdtr', we have 
sthdtur, p. -i%h, in the phrase sthdtup cardtham, a loosely formalized 
expression for 'all beings.' It occurs as N. in i.58.5, sthdttip 
cardtham bhayate patatrinah, where construction and meaning 
are clear. In i.68.1, it is A.s. In i.70.7, vdrdhdn yam pUrvt'h 
kshapd virilpd stfidttip ca rdtham, the plural vdrdhdn may be 
construed ad sensttm with the subject sthdttig, 'cardtham (cf. sdm 
drata, p. 373 top): 'Whom through many nights and mornings 
(kshapd virUpas, as A.p.f.) all beings worship.' BR. refer the 
word sthdtur of these three passages to a stem sthdti'ir, n. 

4. As for the form -tdri, it certainly stands in the text, and in 
such syntactical relations as do not well admit of a locative, but 
render its interpretation as nominative or accusative pretty certain. 
To this may be added the probability that it is a phonetic repre- 
sentative of the organic form in -tar or -tr. According to the 
Praticakhyas, namely, the vowel r has a r in it, and the r is in the 
middle. Thus the Rik Pr. says : repho ''sty rkdre . . . madhye 
sah, xiii.14. Weber interprets the corresponding rule of the Vaj. 
Pr., iv.145, so that r=.\a-\-^r-\-\a, and he compares the r with 
the Zend erZ. The sound that precedes the r, as well as that 
which follows it, is probably the obscure (samvrtu) a, the neutral 
vowel of the English words org&n (d), but (u). See the notes to 
Ath. Pr. i.36,37. Now I think that the diaskeuasts have taken 
account of this dissyllabic pronunciation of r in four instances (all 
at the end of a pdda), and have written it ari. 

In the first, sd rnacid rnayd' brdhmanas pdtir druhd hantd' 
niahd rtdsya dhartdri, ii._3.l7, the construction demands a 
nominative in apposition with hantd' — ' the support of mighty 
truth' (or ' right'). In ix. 86.42, an accusative is certain syntac- 
tically, whatever the exegetical difficulties may be ; Gr. translates 
antdr iyate \ ndrd ca pdnsam daiviam ca dhartdri thus : ' Goeth 
between (both, the praise of men and the support of the gods, 



Neuters.] Noun- Inflection in the Veda. 423 

i. e.) earth and heaven.' In ix.47.4, svaydm kav'ir vidhartdri 
viprdya rdtnarn ichati, the sense demands an infinitive, and that 
rather in an accusative than a locative relation. I translate : 
' The Wise One himself desires to mete out treasures to the singer.' 
Likewise in viii.59.2, vidhartdri hdstdya vdjrah prdti dhdyi 
dar$atdh: 'The wondrous thunderbolt was put in (Indra's) hand, 
to wield it.' 

Dative Singular Neuter. 

No example. In i. 164 15, sthdtre, which Say. takes as dative of 
sthdtdr, is L.s.n. of sthdtrd. 

Genitive Singular Neuter. 

The only example is the word sthdttir, which occurs i.159.3 : 
ii.31.5 : iv.53.6 : vi.50.7 : x.63.8 : and vii.60.2 (vipvasya sthdtur 
jdgatap ca gopd'h). 

Locative Singular Neuter. 

The only example is found in v.9.5, with lengthened final as 
antepenultimate of a pdda of 8 syllables, pipite dhmdtdri yathd, 
p. -ri. 

Genitive Plural Neuter. 

In i.70.3, the text has gdrbhap ca sthdtd'm gdrbhap cardthdm. 
BE. suggest sthdtrd' rh. Sthdtrd' m is to sthdtrnd'm as svdsrdm 
is to svdsrndm. See p. 430, G.p.f. Gr. refers the form to the stem 
sthd'nt, present participle of sthd; cf. pd'nt-am, bhdt-i'. 

THE MASCULINES AND PEMININES. 
The neuters being thus enumerated, we may proceed to the 
masculines and feminines ; and since these are declined precisely 
alike, except in the A. p., they may be treated together. 

Nominative Singular Masculine and Feminine. 

The form ends in -d. It occurs 973 times. Of these occur- 
rences, 870 are masculines (from 140 stems), and 103 are feminines 
(from 6 stems). 

Examples of the masculines are : avitd', 38 ; janitd', 26 ; jaritd', 
17; trdtd', 13; tvdshtd, 45; ddtd', 13; dhartd', 13; neid\ 13; 
samtd, 22 ; savitd', 107 ; stotd', 12 ; hdtd, 158 ; further, pitd', 85 ; 
bhra td, 10 ; trimdtd'; dvimdtd', 4 ; saptdsvasd ; hatdbhrdtd, 
hatamdtd, hatdsvasd, AY. ii.32.4. 

The feminines are : abhrdtd' ; duhitd', 29 ; mdtd', 63 ; saptd- 
svasd; sindlvumdtd; svdsd, 8. 

The analogy of the related languages indicates that this final -a stands for -dr 
Inis m turn must have been developed out of an original form -ar-s (cf. dtar-s, 
Yacna, lxn.7, etc.). The steps of this development are traced by Curtius, Siudien, 

Grassmann, in the preface to his Wb, p. vii., observes that crasis, as shown by the 
metre, always takes place between final -a or -a and an initial r-, when concurrent 
in the mtenor of a pada, except as stated below, and that it results in ar: thus 
mdra somam prta ytuf, 1.15 1; yde cid dhi vdm purd r'shayo, viii.8.6, p. purd' 
(i-rat. ii.ll). Crasis does not take place, i.e., hiatus ensues, (]) when the con- 



424 G R. Lanmun, [Stems in r. 

current vowels are separated by the caesura: as in iv.23.7c; (2) when the r- is 
followed by a double consonant : as in iii.41.2a ; and (3) when -a or -d stands for 
■as, -ds, -e, -ai, -dn, or -dr: thus, indro brahmd' indra r'shih, viii.16.7: kanvd 
rtasya dhd'rayd, viii.6.8. Accordingly, when the metre shows hiatus between -a 
(=-«s) or-d (=-<2s) and r-, we may assume that at the time when the hymns were 
reduced to written form, the endings -as and -ds had not yet lost their final s. In 
like manner, hiatus between -d (as ending of N.s. of -tar) and r- would indicate 
that the historical predecessor of -d, namely, -dr. had not entirely gone out of use, 
or had at least left its graphical reflex in the text. 

To put this matter to a thorough test, I examined every one of these 973 forms 
as it stands in the sarhhitd, and obtained the results here given. In the interior 
of a pdda, if the final -d of the N.s.m. and f. is followed by a vowel, the two are 
almost invariably combined. Thus -a unites with d- : as in x.4.5c ; with e- to at, 
in ix.73.3; with a-, very often; with i- (especially the i of iva) to e; and with u- 
to o : thus, janitd'gner janitd' sil'riasya janitendrasya janitota vishnoh, ix.96.5. The 
form appears before a /vowel with hiatus in only 16 instances: i.60.46; 61.46; 
186.6a: iii.54.126: (in iv.6.2c, read bhaanum?) v.46.4d (tvdshtd- uta) : vi.23.4a; 
24.5d; 25.76: vii.40.3a": viii.19.26c : x.6.15; 49.106: 60.7a (ayam mdtd', aydm 
pitd'); 61.9c; 99.3a; 132.6a. 

The form in -d was followed by r- in only five passages. These, 
along with a sixth one from the Atharvan, follow : 

ii.28.4a. pra stm ddityo asrjad vidhartd'fi (r-), p. -td' 

v.45.6. apa yd' mdtd'fl rnuta vrajaih goh, p. mdtd' 

i.127.10. jA'rnir hota rshUnaam, p. hotd 

iv.33.5. tvdshtti rbhavas tat panayad vaco vah, p. -td 

v.46.1. vidvd'n pathah pura-etti,' rjti, neshati, p. -td' 

vi.133.4. svasa r'sMndm bhAtakr'tdm bdbh&'va, p. svdsd. 

The metre shows that in the first three verses the concurrent 
vowels are to be pronounced with hiatus (-d r-) ; and that in the 
last three they are to be fused to ar and pronounced as one sylla- 
ble. According to the Rik Pr., ii.ll, -d becomes -d, before r-; and 
the metre — as well as the actual prescriptions of the TPr. (x.8) 
and Ath. Pr. (iii.46) — requires also the conversion of the r- to r-. 
That this shortening does not take place in ii.28.4 and v.45.6 is 
plain evidence that the diaskeuasts took due account of the 
metrical value of the -td r- as two syllables; and their way of 
writing them is expressly mentioned by the Prat., ii.31. Why 
have not the diaskeuasts, and after them Qauiiaka, treated the 
third passage, which is precisely parallel, in the same way, and 
written hdtdn rsh-dnd'm ? Kuhn, JBeitrage, iv.212,213, q. v., is of 
the opinion that these hymns of Paruchepa, especially i.l2V and 
129, are comparatively recent productions and belong to a later 
redaction. This singular discrepancy (hdtd r-) is, as I think, satis- 
factorily accounted for by this assumption, and is also a beautiful 
confirmation of Professor Kuhn's criticism. 

Paruchepa is indeed mentioned by Qaunaka, ii 32, in connection with words 
oxurring in i.129 and 133. 

The second and third passages are the ones cited by Gr. : in 
them he would restore the old forms mdtd'r and hdtdr; and if we 
allow these, we may add vidhartd'r. Cf. Delbrflck, Verbum, p. 50, 
and Leskien, Die Declination im Slavisch-litauischen nnd Ger- 
manischen, p. 24. 



N.s.m.f.] Noun- Inflection in the Veda. 425 

As for the last three passages, on the other hand, it seems not 
unlikely that in the time of the oral tradition combinations like 
ddgdhd'si (for *ddgdhdr dsi, v.9.4) and tvdshtarbhavas (for *tvdsh- 
tdr rbhavas) stood quite on a level with svddhittva (for svddhitir 
iva, p. 375 end) in respect to elision and crasis, and with avitd' 
rdthdndm (for -td'r r-) in respect to elision. 

In iz.9T.38, we have sa pundnd upa sii're nd dhd'tdbhe aprd rodasi vi sha dvah, 
p. dhd'td d' ubhe. The combination is interesting as showing how entirely pre- 
posterous and artificial is the sandhi between two pddas. The d' belongs of 
course to pdda b, and we must pronounce obhe aprd, etc. The graphical combina- 
tion of dhd'td with obhe would give dhd'taubhe ; but instead of this, we have dhd'td 
first combined with d' and afterwards with ubhe. Compare, however, Prat, ii.31, 
r. 61, clxvi. 

Accusative Singular Masculine and Feminine. 

Masculine. A. The form ends in -tdram : as, ydtd'ram. It 
occurs 168 times (from 41 stems). 

Enumeration : adhivaMd'rarn ; avasdtd'rarn ; avitd'ram, 3 ; astdram ; dyan- 
td'ram; ishkartd'ram, 2; kartdram; gdntdram, i.9.9 ; cettdram ; jaritd'ram, 11 ; 
jetdram, 3 : tarutd'ram, 3 : trdtd'ram, 5 ; tvdshidram, 9 ; ddtd'ram, 3 ; dd'tdram, 
iv.31.T ; dhartd'ram, 3; niddtd'ram; praneld'ram, viii.16.10; netd'ram; pani- 
td'ram ; pavitd'ram ; prahetd'ram; bodhayitd'ram; mandhdtd'ram ; marditd'ram, 
5; yantd'ram; ydntdram; ydtd'ram; rakshitd'ram; vanditd'ram; vibhaktd'ram; 
frotdram; sanitd'ram: sarneddhd'ram, 2; s avitd'ram, 11 ; stotd'rarn, 12; han- 
td'ram, 2; haskartd'ram, iv.T.3; hetdram ; hotdram, 62. 

B. Here belong 57 forms (from 7 stems) : ddrimdtaram ; jd'nid- 
taram; devdram; ndram, 16 ; pitdram, 33 ; bhrd'taram, 4 ; sin- 
dhumdtaram. 

Feminine. A. Sole example : svdsdram, 5 ; AV. i.28.4: iii.30.3. 

B. Here belong 31 forms (from 3 steins): duhitdram, 3; md- 
tdram, 27; saptdmdtaram. 

Instrumental Singular Masculine and Feminine. 

Masculine. Here belong : dstrd ; dhdtrd' ; ndptrd, 2 ; pitrd', 
2; savitrd', 2; from the AV., dstrd, xi.2.7; tvdshtrd, xii.3.33; 
bhrd'trd, v.22.12. 

Feminine. Here belong : duhitrd' '; from the AV., mdtrd', iii. 
30.2; svdsrd, v.22.12. 

Dative Singular Masculine and Feminine. 

Masculine. Here belong 52 forms (from 7 stems) : jaritr'e,, 23 
ddtr'e; ndptre, 3 ; pitre, 9; savitre, 2 ; stotre, 13 ( Wb., col. 1 774) 
hdtre; from the AV., dstre, vi.93.2 ; kartre, x.1.30; kroshtre, x\ 
2.2; jaritr'e, xviii.1.40; ddtr'e, ix.3.12 ; 4.1: x.9. 13-24 ; 10.27 
dhdtr'e and vidhdtr'e, iii.10.10 ; pitre, i.31.4 : v. 11.1 : vii.14.3 
rakshitre, xii.3.55-60. 

Feminine. Here belong: duhitre ; rndtr'e ; svdsre ; from the 
AV., duhitre, vi.137.1 ; rndtr'e, i.31.4. 

C-stems. Here belong: ndre, 5 ; st'iarnare. In iii. 18.1, pitdreva, 
p. pitdrd-iva, Bollensen sees a dative, and divides thus: pitdre va. 
There is nothing improbable in the form as dative ; but the sense 
favors decidedly the view of the padakdra. 



426 0. R. Lanman, [Stems in r. 

Ablative Singular Masculine and Feminine. 

Masculine. Here belong : dstur; dhdttir, 3 ; pitur, 5 ; bhrd'tur; 
vijdmdtur ; savitur, 3; hdtur, 2; from the AV, tvdshtur and 
dhdtur, xi.8.9 ; pitdr, iii.25.5 : vi.116.3; bhrd'tur, vi. 116.3. 

Feminine. Here belong : duhitur; mdt'ur, 6 ; svdsur, 3 ; from 
the AV., mdtur, iii.25.5 : vi.116.2,3: viii.9.5. 

Genitive Singular Masculine and Feminine. 

Masculine. Here belong 129 forms (from 24 stems) : abhi- 
kshatt'&r; avitur; dstur, 3; janitur, 3; jurit'dr, 19; trdtur, 2; 
tvdshtur, 8 ; ddti'ir; ndptur, 2 ; nidhdtur; netur; neshtur; pitur, fc 
36 ; bhrd'tur, 3 ; mandhdtur; vanditdr; vdvd'tur, 2 ; vddhur; 
pamitur; sanitur; savitur, 29 ; sotfar; stotur, 4 ; hotxir, 6 ; from 
the AV., utthdtur, ix.4.14; cetttir, vi.73.1 ; dhdt-dr, ii.36.2 ; trdttir, 
etc. 

Feminine. Here belong : duhittir, 7 ; mdtdr, 22 ; svdsur, 2 ; 
from AV., ndndndur, xiv. 1.44 ; mdtur, 6 times ; svdsur, xviii.1.14. 

C-stems. The full ending -as with the correspondingly weak 
theme is seen only in usr-ds (for ushr'-as, stem ushdr), iii.58.4 : 
vi.12.4; but the monosyllabic theme ndr makes ndr-as, i. 12 1.2 : 
v.9.7 : vii.31.2. 

I am unable to find any support for such a form as pitr-ds, cited by Benfey, 
Gram. p. 300, note 8. It is hard to say whether the form is to be set up as end- 
ing in -ur or -us. Schleicher, Comp. 4 p. 538, §252, traces the genesis of the form 
thus : *pitar-as, (*iypitrds, *pitr's, pitus. The analogy of ve-s as compared with 
avy-as, and of madho-s as compared with mddhu-as, however, favors our starting 
with *pitdr-s (or pitr'-as ; but not with pitar-as) ; cf . nar-s, Yc. iii. 1 7. The series 
*pitdr-s, *pitdr-r, *pitr'-r, pitur seems to me perfectly licit and simple ; but to 
some, in view of the analogy of the Pali usabha, vusa (=rshabha. vrsha — Kuhn, 
Pdli-Gr. p. 14), the following may appear more probable: *pitdr-s, *pitr'-s, pitus. 

The Tedic n-dr-as—we should expect nur, for *ndr-s — is perhaps to be put on 
a level with girdy-as, p. 383, as having the fuller ending along with the stronger 
form of the stem. See p. 420 med. 

Locative Singular Masculine and Feminine. 

Masculine. 1. Here belong : netdri; pitdri; sotdri, 2 ; vaktdri, 
AV. ii. 1.4. For ndri, i.85.9 : viii.85.19, see p. 348 med. 

2. The samhitd has -i, and the pada reads -i. The protraction 
is in every instance metrical. Here belong; kartdri, i.139.7 (i is 
tenth syllable in a pdda of 12); vaktdri, x.61.12 (eighth of 11) ; 
dhmdtart (n., see p. 423). 

In v.41.10 and vi.12.4, etdri nd fashaih, p. etdrl, the i is eighth in a pdda of 11. 
G-r. takes etdri as fem. of etdr('wie die Wallerin mit Liedern'); cf. avitdri, vii.96.2, 
text -tri'. The padaMra's interpretation as L. rests, perhaps, on the analogy of 
the cases just given. I find no mention of them in the Pra.t., although they are 
co/ored by the general prescriptions of viii.21,22. 

Feminine. Here belong: duhitdri, i.71.5 (i, being followed by 
a double consonant, is not lengthened, although the eighth in a 
pdda of 11 syllables) ; cf. p. 333 ; mdtdri, twice, and AV. xi.7.6. 

C-stems. In x.85.46, ndndndari samrd'jni bhava, we must pro- 
nounce ndndndri. The Atharvan gets over the difficulty by 
reading ndndnduh. 



L.s.m.f.] Noun- Inflection in the Veda. 427 

Contrariwise, in v.53.14, vrshtvi' pdm y6r a! pa usri bheshajdm, 
we must restore the syncopated vowel and read ushdri. 

In x.6.5, usrd'm seems to be an isolated ease of transition to the 
2-declension (series B, p. 366 end). Pronounce ushar-d'm. 

Vocative Singular Masculine and Feminine. 

The form is identical with the simple stem. There are HO m. 
forms (from 26 stems), and 20 f. forms (from 3 stems). 

Masculine. Here belong : avaspartar; aoitar; janitor; jaritar, 
1 1 ; jdmdtar ; trdtar, 3 ; tvashtar, 4 ; dartar / doshdvastar, 3 ; 
dhartar; dhdtar, 2 ; netar, 2 ; neshtar; pranetar, 5 ; prayantar ; 
yajftahotar; vidhartar; vidhdtar; sanitar; savitar,2S; susanitar, 
2 ; sotar; sthdtar, 5 ; hotar, 27 ; further, pitar, 4 ; bhrdtar, 6 ; 
from AV., vipastar, ix.5.4. 

Feminine. Here belong: duhitar, 17; mdtar, 2; from AV., 
duhitar, xix.47 .5 ; mdtar, 48.2 ; xii.1.63. 

C-stem. In i.49.4, ushar (Prat, iv.13) is V.s. of ushdr. 

Nom., Acc, and Voc. Dual Masculine and Fekinine. 

The ending -d is used 176 times; the ending -au, only 10 times. 
The circumstances of occurrence coincide entirely with those given 
on p. 340, q. v. 

I. The ending -d is used : 

1. at the end of apdda (51 times) ; 

2. before consonants (112 times) ; 

3. before an initial vowel with which it is fused (11 times). 

4. It is never used before vowels with hiatus. 
II. The ending -au is used : 

I . before vowels without hiatus, in the form -dv (8 times). 

_ I. 1. In vii.94.3 : viii.8.17 : x.40.4, the following pdda begins with i- ■ and in 
vi.60.9: viii.26.11 : x.]43.6, with »-. The concurrent vowels are written as 
fused ; but are of course to be pronounced with hiatus. 

3. The final -d coalesces with a- in i. 110.8 : x.39.3 ; with a-, in V.73.Y • with 
iva, in iii.18.1 ; 58.2: iv.41.7 : ix.18.5: x.59.1; 106.2,4; with i-, i.117.18. 

4. In i. 161. 10 and x.65.10, the samhitd has -a u- in the interior of a pdda, and 
the pada has -au u-. Cf. p. 341, 4a, and Hik Pr. ii.9, and Ath. Pr. iii.40. 

II. 1. There are only two instances of -au before a consonant : viz., at the end 
of iii.55.11c (before tn-), and of x.U.lla (before o). iii.55 is one of the later, 
mystical hymns ; and verse 11 of x.14 contains notions about the hounds of Yama 
which are entirely different from those of verse 10, and probably later. The 
enumerations follow : 

I. Masculine. A. Here belong 31 forms (from 15 stems). They 
are: avitd'rd, 2; asndtd'rd; itshtd'rd ; gdntdrd, 4; coditd'rd; 
janitd'rd; dhartd'rd, 3 ; nlcetdrd; pretd'rd; yantd'rd; -yantd'rd; 
rakshitd'rd; pamitd'rd ; sthdtd'rd, 2; hotdrd, 10; from AV., 
dhartdrd, vii.73.4. 

B. C. Here belong 109 forms (from 9 stems). They are : mdtdrd- 
pitdrd ; ihehamdtard ; ddkshapitard ; pitdrd, 23 ; bhrd'tard • 
sindhumdtard ; (ubhd 1 7) mdtdrd, i.140.3; nard, 69; ndrd, 9^ 
finally, manotd'rd, 2 : see p. 420 sub fin. 

Feminine. A. Here belongs svdsdrd, 3. 



428 C. R. Lanman, [Stems in r. 

B. Here belong 33 forms (from 3 stems). They are duhitdrd, 
2 ; mdtdrd, 30 ; sammdtdrd; from AV., mdtdrd, v.1.4. 

II. Masculine. A. Here belong: ddtd'rau; rakshitd'rau. The 
AV. has: anushthdtd'rau, xv. 4.1-6 ; Jcshattd'rau, iii.24.7 ; gop- 
td'rau,v.30.\0: xv.4.1-6. 

B. C. Here belong : pitdrau, 3 ; narau, 2 ; dhdnutdrau : see p. 
420. The. AV. has: pitdrau, vi.120.3: xiv.2.37 ; satnmdtdrau, 
xiii.2.13. 

Feminine. A. Here belongs sudsdrau, iii.55. 11 : AV. v.2.9. 

B. Here belong: mdtdrau; from AV., duhitdrau, vii.12.1. 

Inst., Dat., and Abl. Dual Masculine and Feminine. 
The only example is pitr'bhydm, thrice^ as dative. 

Genitive and Locative Dual Masculine and Feminine. 

The form occurs 25 times (from 3 stems). It is always written 
-tros (sros) ; but the metre shows that the word is to be pro- 
nounced as a trisyllable (pi-tr-6s, etc.) in every instance (except 
vii.3.9c), and so also in the single form from the AV. See p. 420 
med. 

Genitives: pitrds, 14; mdtrds, iii.2.2 ; pitrds, AV. xx.34.16. 

Locatives : pitrds, 6 ; mdtrds, 3 ; svdsros. 

Nominative and Vocative Plueal Masculine and Feminine. 

Masculine. A. Here belong 69 forms (from 40 stems). 

Enumeration : agnihotdras ; abhikshattd'ras ; abhisvartd'ras ; dsldras ; upalcshe- 
td'ras; gantdras ; cetd'ras (ptcp.) ; cetd'ras ; jaritd'ras, 10; joshtd'ras ; td'ras ; 
trdtd'ras, 2 ; ddtd'ras ; dhartd'ras, 2 ; dhd'tdras ; nicetd'ras, 2 ; ninditd'ras ; 
netdras; netd'ras, 2; nrpdtd'ras ; panitd'ras, 2; pavitd'ras ; prajndtd'ras ; prane- 
td'ras; prdvitd'ras; pretd'ras; yantd'ras; rahshitd'ras ; vantd'ras, 2; vidhdtd'ras; 
viyotd'ras ; famitd'ras ; frotdras ; sanitd'ras ; sotti'ras, 4 ; stotd'ras, 1 ; sthdtd'ras ; 
svaritdras ; hetdras ; hotdras, 6. 

In i.62.3 : ii. 19.1 : v.33.5, the first a of ndras appears to have 
the value of a long syllable, as penultimate of a trishtubh-pdda : 
thus, nd'ras. 

B. C. Here belong 223 forms (from 9 stems). They are : gomd- 
taras; ddkshapitaras; ndras, 156 ; pitdras, 48 ; pr'pnimdtaras, 
10 ; bhrd'tarus, 4 ; sindhumdtaras; sumdtdras; snarnaras. The 
AV. has: devdras, xiv. 1.39; vasudhd't&ras, v.27.6 (see p. 420 end.) 

Feminine. A. Here belongs svdsdras, 21. 

B. Here belong : abhrdtdras; duhitdras, 3; mdtdras, 23; from 
A V, abhrd'taras, i.17.1 ; duhitaras, ii.14.2 ; mdtdras, vi.9.3 : xix. 
40.3 ; sammdtdras, viii.7.27. 

If the form str-n-as — cited as N.p. from Jyotisham, pp. 4, 52 — is 
authentic, it must be regarded as a form of transition to the 
w-declension ; cf. khddi-nas, p. 393. 

Accusative Plural Masculine and Feminine. 

Masculine. The usual ^a<fa-form -fn occurs 65 times (from 12 
stems). 1. In the interior of a pdda, it appears in the samhitd : 



A.p.rn.f.] Noun-Inflection in the Veda. 429 

(a) as -fnp before cy-, x.50.4 (Prat, iv.32) ; as -fnr before a-, v.54.15 
(Pr. iv.30) ; as -fnh before p-, i.121.1 : viii.73.3 (Pr. iv.34) ; (b) as 
-fn, 16 times before consonants : e. g., x.64.8; 154.4; as -fn, before 
a-, i.42.5 : ii.15.5 ; 34.14 : v.50.3 ; before i-, vii.55.3 ; before r-, 
vi.39.5; (c) as -fn, before,/-, vii.3.8 ; and before c-, i.72.6 (Prat. 
iv.32 ; but see p. 422, 2). 2. At the end of a pdda it appears as 
-fn in 36 instances : e. g., v.33.1, and as -fn before j- in v.15.2. 

The forms are : dstfn; asndtf'n; jaritf'n, 2 ; trdtf'n; ddksha- 
pitfn; ddtf'n; nf'n, 42 ; pitf'n, 7 ; pravodhf'n; stotf'n, 3 ; sthd- 
tf'n; hdtfn, 3. The AV. has: kartf'n, x.1.14,17; goptf'n, xix. 
27.4; pdtfn (a-), iii.12.8 ; pitf'n, 13 times; pitf'nr (w-), xviii.2.4, 
23 ; 4.40 ;' pr'pnimdtfns (t-), iv.27.2. 

In the A.p. the stem would have its weak form : as, pit? ; this, with the usual 
case-ending of vocalic stems, -ns, and the lengthening, gives pitf'ns. Of this 
organic form, the Pr&t, iv.30,32,34, gives four examples from the Rik, and two 
others taken from Praishas (nrfili patibhyah and nrHh pranetram). The treatment 
of the form before vowels ought d priori to be the same as that of the forms in 
-dn, -in, -An (p. 346, 395, 415); but in fact it is quite arbitrary, as may be seen 
above. See also note to Ath. Pr. ii.29. 

Schleicher thought that the original form consisted of a consonantal stem with 
corresponding ending : as, pitdr-as ; but this seems to me unlikely for the period 
of separate Indian development. The metre, however, suggests the reading nar-as 
in vii.28.3, text nf'n, and x.50.4, text nf'iif. In x.35.106, Gr. reads hdtaras, text 
hdtfn; but the verse may consist of 12,11,12,12. 

Feminine. Here belong: rndtf's, 4; svdsfs,2. Inx.35.2,mdtf'n 
is declined as a masculine! Cf. i.140.3. We see the ending 
which belongs to consonantal stems in only one word, usrds, 
occurring thrice. 

INSTBUMENTAL PHJEAL MASCULINE AND FEMININE. 

For the remaining cases, the ending is added to the weaker 
form of the stem. 

Masculine. Here belong 112 forms (from 14 stems). They are: 
dstrbhis; kartr'bhis, 3 ; dhdtr'bhis, 3 ; ndptrbhis; partr'bhis, 2 ; 
pitr'bhis, 9; saptdmdtrbhis ; setr'bhis ; sotr'bhis, 7; sdtrbhis ; 
hetr'bhis, 2 ; hdtrbhis, 2 ; further", nr'bhis, 71 ; str'bhis, 8. The 
AV. has : nr'bhis, 2 ; pitr'bhis, 5 ; bhrd'trbhis, vi.4.1. 

Feminine. Here belong : mdtr'bhis, 6 ; svdsrbhis, 2. 

Dative Plural Masculine and Feminine. 

Masculine. Here belong 70 forms (from 4 stems). They are : 
jaritrfbhyas, 11; nr'bhyas, 18; pitr'bhyas, 9; stotr'bhyas, 30. In 
i.110.6 and x.148.4, Gr. reads nr'bhias; but this is not certainly 
necessary. The AV. has: pitr'bhyas, 17; rakshitr'bhyas,iii.27.1-6. 

Feminine. Here belongs mdtr'bhyas, i.95.7. 

Ablative Plural Masculine and Feminine. 

Masculine. Here belong: nr'bhyas, i.173.6 : vii.20.5; pitr'bhyas, 
twice. 

Feminine. Here belongs mdtr'bhyas, x.1.2. 
vol. x. 58 



430 C. R. Lanman, [Stems in 

Genitive Plural Masculine and Feminine. 

Oxytone stems in r', like those in i and -& (p. 397, 417), shift 
the accent to the ending in the G.p. In accordance with the 
analogy of the i and w-stems also, the r should always be length- 
ened. It is always written as long in the Rik, except in the case 
of nrnd'm. 

Masculine. I. The thematic f is long. 1. Here belong : dhdtf- 
nd'm; pitrnd'm, 6 ; stotfnd'm, 3 ; hdtfndm (the only barytone). 
The AV. has : ddtfnd'm (MSS., ddirend'm ; ed., ddtrdnd'm), 
v.24.3; pitfnd'm, 11 times: as, ii.12.4; so TBr. ii.6.16 2 . 

2. Gr. reads -aarn where the text has -dm at the end of a pdda 
of 7 syllables: jaritrnd'm, i.30.15 : iv.31.3 : vii.66.3 ; pitrnd'm, 
vi.46.12. 

3. Resolution is necessary in 5 instances : pitfndam, viii.41.2 : 
x.57.3; stotfndam, v.64.4 : vi.45.29; hdtfnaam, viii.91.10. 

I. a. Benfey, Vedica, p. 1-38, has shown clearly that in the 
Veda r designates not only the short, but also the long vowel. In 
iii.52.8: iv.25.4 : v.30.12 : vii.1.11 ; 19.10; 62.4: x.29.2 ; 99.9; 
111.1, nrnd'm stands at the end of a trishtubh-pdda, and the 
syllable nr as penultimate has of course the value of a long 
{nrnd'm). Ini.48.4: iii.16.4: v.18.5: vii.32.11: viii.55.5, nrnd'm 
stands at the end of a pdda of 7 syllables (Gr., reading -aam, 
makes 8), and must be pronounced nrnd'm; so also at the end of 
x.93.125 (11 syllables — Gr., -aam, 12). 

II. The thematic r is short. 1. Here belongs nrnd'm, i.43.7 : 
ii.1.1: iii.51.4: vi.65.5: vii.83.7. 

3. Resolution is necessary (nrndam) in i.7 7.4: vi.33.3: viii.40.2 : 
x.29.1; 93.4; 148.4. 

The AV. has nrnd'm, ii.9.2 : xix.47.7. The TS. has : udgdtr- 
nd'm, iii.2.9 6 ; dhdtrnd'm, iv.7.14 3 (Rik x.128.7, -f-) ; netrnd'm, 
i.3.6 1 ; pitrnd'm, i.3.6 1 ; 8.5" (Rik x.57.3 : VS. iii.53, -r-) : iii.3.5 1 
(so Bhag. P. iv.15.8) ; bhrd'trndm, ii.6.6" ; so strndm, Jyotisham, 
p. 89. See Weber, Jnd. Stud, xiii.101. 

C. The only stem which is treated as consonantal in the G.p.m. 
is ndr. This makes nar-d'm, 6 ; nar-dam, 10. 

Ludwig, Der Infinitiv im, Veda, p. 6, takes nf'n as a genitive in i.121.1: v.7.10: 
i.181.8: iii.14.4: iv.2.15: v.33.1 : vi.2.11 ; 3.6: x.29.4: i.146.4. Inthefirsttwo 
passages, the text is corrupt ; in the last, Gr. also takes it as genitive (see Veb. 
ii.508), and Roth thinks this interpretation possible. If genitive, it stands for 
nf'm (?); cf. y&thiflm, etc., p. 353. 

Feminine. Here belongs svdsfndm, i. 124. 9: iii.1.3,11. 

C. The only stem, not monosyllabic, to which the ending is 
added directly, is svdsar — svdsr-dm, i.65.7. I consider this a 
wrong formation, to which, perhaps, Paracara was forced by the 
exigencies of his favorite metre. Schleicher would regard it as a 
relic of the older mode of formation. Cf. usrd'm, L.s.f., and see 
p. 423, G.p.n. 

Locative Plural Masculine and Feminine. 

Masculine. Here belong : devr'shu; nr'shu, 2 ; Mtrshu. The 
AV. has pitr'shu 13 times : as, i. 14. 1,3. 

Feminine. Here belongs mdtfshu, 4. 



ai, o, au.~] Noun- Inflection in the Veda. 431 



STEMS IN AI, 0, AU. 

Under this head belong only five simple stems and their com- 
pounds : rai, rdhddrai, brhddrai ; go", dgo, pr'$nigo, r&padgo; 
dy6, prthivi'dyd, pradyd, vrshtidyo; nau; glau. They are used 
only in the masculine and feminine, and between these two gen- 
ders there are no declensional differences. Although the thematic 
vowels are not suffixal, I have nevertheless followed the tradi- 
tional order, and treated these stems after those in r. 

I. The stems rai (m.f.) and ray'i (m.f.) supplement each other 
as follows : 

N. rayis, 28 rdfyas, 20 

A. rayim, 180 rdyds, 22 

I. rdyd', 39 rayibhis, i.64.20 

D. rdye, 62 

Ab. rdyas, 3 

G. rdyds, 103 raytnd'm, 41. 

The stem ray'i intrudes twice on the domain of rai in the I.s., 
forming rayyd' , x.19.7, and rayind, x.122.3. 

The stem rai intrudes twice on the domain of rayi, forming 
rd'm as A.s. at x.111.7, and — perhaps for the sake of the metre — 
rdyd'm as G.p. at ix.108.13 (Gr. unnecessarily, -ddm). 

In i. 129.9,10, rdyd' seems to be trisyllabic {rdidl or rayind f). 
In i.68.10 : iv.41.10 : vi.19.5, the G.s. appears with irregular 
accent, rd'yas. So in vi.47.9 : vii.34.22 ; x.140.4 : Val. 4.10, the 
A.p. appears with irregular accent, unless indeed we assume that 
in the matter of accent the A.p. is -treated as a strong case (cf. 
vr't-as, bhid-as, etc.). 

In the Atharvan, rdyd' occurs at xiv.2.36 ; rdy'e, xviii.2.37 ; 
rdyds (sing.), 22 times. 

The Rik has three forms which seem to be compounds of rai : 
brhddr&ye, rdhddrdyas, and patd-rd. The first two are referred 
by BR. to stems in -ri, -ri. This is possible (see p. 384, G.s.m.) ; 
but it seems to me more likely that the forms stand for -rdye, 
-rdyas, the & being due to the metre. See p. 382, D.s.m. The 
word patdrd, x. 106.5, we may regard as coming from the more 
original form of the stem (-rd) ; or, possibly, as standing for 
patard'?. 

The noun-stem would seem to have been once identical in form with the verbal 
root rd ; and we have the forms rd-m, rd-bhydm, rd-bhis. The y is perhaps of 
phonetic character (as in devd-y-os). The grammarians abstract the stem rd'y or 
rai from the form rd'-y-as, by dividing it as rd'y-as. 



II. The stem g6 is thus declined : gaits, 22 ; A., gd'vam, text 
gd'm, 6: gd'm, 34; gdvd; gdve, 13; Ab., g6s; G., gdvas, text 
g6s, 3 : g6s, 34 ; gdvi, 1 ; gd'vd, 2 : gd'vau, 4 ; gd'vas, 97 : gdvas, 
3; A., gdvas?, text gd's, i.61.10: gd's, 100 times, and 458,6; 
gdbhis, 76 ; D., gdbhyas, 2 ; G, gdvdm, 55 : gdndm, 15 : gdndm 
(at end of pdda of 7 syllables — Gr., -aam), 5 ; gdshu, 43. 

The dissyllabic character of gd'm (=gd'uam) is clear in v.52.16, 
where gd'm is metrically parallel to pr'pnim : gd'm vdeanta surd- 



432 0. B. Lanman, [Stems in 

yah | pr'pnim vocanta mdtdram. So viii.4.21. BR. cite gdvas 
(d!) as A. p., from TBr., etc. Gdndm is never used except 
at the end of a pdda. See Pan. vii.1.57 ; gdndm — cited by Boht- 
lingk, ad loc, as an exception — is at the end of an aksharaparikti, 
i.69.3c. Gdvdm is regularly used in the interior of a pdda (52 
times : in v.30.4, pronounce tirudm — not gdvadrn). It is used 
twice at the end of pddas with an even number of syllables : 
x.166.1 (8 syll.): ix.81.1 (12 syll.). For iv.l.l9c (10 syll.), Gr. 
suggests gdndm; but the metre is still short (read p&ci u-'dho [(J'] 
atrnan nd gdndm ?). 

In the AV., the metre gives no evidence of the forms gd'vam 
(A.S.), gdvas (G.s.), gdvas (A.p.) ; gdvd, gdve, and gd'vd are not 
found ; the other forms are like those in the Rik. Gdndm occurs 
xx.127.3 (end otpdda); gdvdm, 11 times (interior of pdda). 

The Rik has the compounds : d-gos, G.s.m. (may come from 
d-gu, p. 401); pr'pnigdvas (cf. ddhrigdvas, p. 414), N.p.m. ; 
r&padgavi, L.s.f. 

IIL The stems div (m.), dih or dy-d, (m.), and did or dyd (m.f.) 
supplement each other as follows : 





A. did, dyd. 




B. div. 0. diii, dyu. 


N. 


diaus, 26 dya&s, 46 






A. 


did' in, 12 dyd'm, 67 






I. 






divd', 9 


D. 






dive, 15 


Ab. 






divas, 50 + 


G. 






div&s,\&0 + 


L. 






divi, 118 


V. 


diaus, 2 dyaus, 4 






N.A.T. 


dyd'vd, 26 






N.V. 


dyd'vas, 22 






A. 






dy&'n, 24; dM'n 


I. 






dyubhit, 19. 


The form dyd'm is read 


fcy 


Gr. as did'm in 12 passages ; 



but this is justified only in 8 or 9. We must read i.l21.2d as a 
catalectic pdda (pdrijmdnam iva dyd'm) so as to keep the a of 
iva long by position. The requirements of the metre are better 
satisfied by tipa dyd'vam skambhdthu skdmbhanena, vi.72.2c, 
than by ■dpd did'm; so in i.67.5c {tastdmbha dyd'vam), and per- 
haps in ii.ll.15dL The form dyd'm is related to the form dyd'vam 
thus rendered probable, as gd'm to gd'vam, and perhaps as rd'm 
to rd'yam. 

The vocative dya&s occurs at viii.89.12 : x.59.8,9,10 — of course, 
as a monosyllable. In vi.51.5 — the only passage in the Rik where 
it occurs as a dissyllable — the diaskeuasts have taken due notice 
of the metrical value of the word, and accented it, in accordance 
with the general rule, on the first syllable, diaus; this, of course, 
appears in the written text as dyaus, with jdtya svarita. In 
AV. vi.4.3, also, we have diaush pUar ydvdya duchtind yd' (ed. 
wrongly, dyaush). Compare the voc. jySke (i. e. jidke, from 
jydka), AV. i.2.2, and vyd"ghra (i. e. vidghra, from vydghrd), 
iv.3.3. The MSS. have the impossible reading vydghra, and the 
edition wrongly vyd'ghra. These have been pointed out by 



at, o, au.~\ Noun- Inflection in the Veda. 433 

Whitney. The circumflex is in each case perfectly regular. 
Quite anomalous is the dual dydv-i, iv.56.5, formed with the 
ending of a neuter. 

The stem dy6 intrudes on the domain of div twice. The Ab.s. 
dyaus, dependent on abhi'ke, i.71.8=TS. i.3.14'=VS. xxxiii.ll, 
is quite isolated. Mahidhara, prathamd shashthyarthe, divah 
samtpe 'ntarikshe. The L.s. dydvi occurs only 12 times in the 
Rik, against 118 occurrences of divi, and only once in the AV., 
where divi is frequent. Dydvi-dyavi occurs twice in the Rik. 
Both dya&s and dydvi are improper formations. 

B. In mandalas i.-vii., divas occurs as Ab. 50 times and as G. 
180 times; in mandalas viii.-x., Ab. and G. together, 167 times; 
besides, divas occurs with a voc. 2 1 times, and divo napdtd twice. 
In the G.d. of dydl vdprthivi' , the first element appears as a singu- 
lar, divdsprthivyds, 4. 

The stem div intrudes on the domain of dyd in the A.s., divam. 
This occurs 21 times (against 79 occurrences of the written 
dyd'm). In a decided majority of these 21 passages may be seen 
other marks betraying a later origin, and I regard divam as an 
improper formation. It is the regular form in the later grammar. 
As a N.p., divas (for dyd'vas), ix.108.11, is hardly admissible. 
Gr. takes it as N.p. in his Wb. ; but as G.s. (accent !) in his Veb. 

The stem div intrudes on the domain of dyH in the A. p., divas, 
ii.3.2 : iv.53.5. In iv.3.8 and v.47.1, it is doubtful whether divds 
(accent !) is an A. p. The only form of div that is certainly plural 
(divas at ii.3.2 and iv.53.5) is feminine. 

C. The stem dyu intrudes on the domain of div and makes 
dyds, occurring as Ab.s. twice, and as G.s. 4 times. The form, 
however, may come from dy6 as well as from dyu, just as dgos 
from dgo or dgu. 

In verses peculiar to the Atharvan are found the forms divd' 
and div'e ; the forms dya&s, dyd'm (e. g. i.2.4), divam, divds, 
divds, and divi occur frequently ; dydvi occurs in xii.2.18 and in 
two Rik passages ; dyu'n and dyabhis, only in Rik verses (xviii. 
1.24 and vi.31.3); divas, as N.p.f., xi.7.14 and xiii.3.21; divas, as 
Ap.f., iv.20.2; 34.4: xix.27.3 ; 32.4 (MSS. divd). 

The following compounds occur : prthivi'dyd'vd ; pradyatis, 
AV. xviii.2.48 ; in the Rik, pradivd; pradivas, Ab., 21 ; pradivi, 
8 ; vrshtidydvd, -dvas; sudivas, N.p. ; dhardivi, once, and AV. 
v.21.6. "We find dyd'vd- in the Rik with -kshd'md (8), -prthivi' 
(79), and -bhDlmi (5). 

Transitions to the a-declension. Here belong the forms divd-m 
(15), dive-dive (46), and tridive", made after the analogy of the 
weak cases div-ds, etc. The AV. has tridivdm, sudivd-, divirdive 
(xx.135.10). 

Adverbially recessive accent is seen in div-d, which occurs 25 
times in the RV. and 13 times in the AV. 

IV. The stem nati, t, is thus declined: naiis ; nd'vam, 17; 
ndva", 13; G., ndvds ; ndvi, 2; N.p., nd'vas, 4; navbhis, 2. 
The AV. has the forms : na<is, 4 ; nd'vam, 6 ; N.p., nd'vas, v. 4.5. 



434 C. R. Lanman, [Rad. d and 

Transition to the <2-declension. From ndv-d' as a stem comes 
the I.s. ndvdyd, i.97.8, sd nah sindM iva ndvdyd. So T.ar. 
vi.ll. 2. Observe the accent. 

V. The stem glati, has the following Vedic forms : glads, AV. 
vi.83.3 ; N.p., gld'vas, Ait. Br. i.25 ; glaubhis, VS. xxv.8. 



STJFFIXLESS STEMS. 

The following section treats of suffixless stems — those whose 
only element, or whose final element, is a root. First come 
the vocalic, and then the consonantal stems, in alphabetical 
order. It will be seen, however, that the root-words ending 
in other vowels than a or d have been already discussed along 
with the stems whose suffixes end in the corresponding vowels. 
This course has been pursued, in order that similar forms 
might be brought together (as, ahias and gdihdnias, p. 384), 
which would otherwise have been widely separated. We 
shall treat of the stems in d and a radical also together, case by 
case. This course has here the same practical advantage as in 
the treatment of the stems in I and 1 u and A 



STEMS IN RADICAL A AND A. 

If we study the mass of forms, we shall see that there are two 
series of declensional endings. The first — which, for convenience, 
we shall call series A — is that which the grammars assign to the 
stems in radical long d, and of which -pd' may serve as "a para- 
digm: 

Series A. : m. and f. 



■pd's 


-pai, 


-pd's* 


■pd'm* 




-pii's 


■pd'* 


-pd'bhydm* 


-pd'Vhis* 


-pe 




-pd'Vhyas' 


-pa's 


-pos 


-pd'm 


-pi 




-pd'su* 


■pds 




-pds* 



Series B : m., 1, and n. 

The other forms coincide entirely with those of the stems end- 
ing in suffixal d or d — ga-td-s, ga-td', ga-td-m, etc. The declen- 
sional series of these stems we shall designate as series B m., B f., 
and B n. 

The Indian grammarians, as is well known, do not admit the 
existence of roots in short d. To this view they were perhaps led 
by the facts that the great majority of the verbal forms actually 
have long d, and that short d at the end of compound nouns often 
supplants a long d of the Veda. In fact, from the Vedic texts 
themselves, it appears that the forms like go-pd'm are, in general, 
younger than those like go-pd'm. 



a-stems.] Noun-Inflection in the Veda. 435 

The great bulk of all the forms from <2-stemB are either nominatives or accusa- 
tives. Of the other cases the examples are few, or even only sporadic. In the 
N.s.m., the Rik-forms with long d are more than ten times as numerous as the 
Rik-forms with short S, and five times as numerous as the Atharvan forms with 
long a: on the other hand, the Atharvan stems with short & are almost as 
numerous as the Atharvan stems with long d, and considerably more numerous 
than the Rik-stems with short S. Moreover, the Atharvan has no masculine 
forms from stems in long d in the oblique cases. It would seem, too, that the 
verses in which the forms with short 8, occur, belong, in general, to the younger 
parts of the Rigveda. 

The later samhitds, as compared with the earlier, show a decided tendency to 
give up the old A-forms and use the new B-forms. Thus in place of the Rik 
ndma-dhd's (x.82.3), the Atharvan variant has ndma-dhd's (ii.1.3). In the Rik, 
we find prathama-jd's and carshani-prd's (N.s.m.), and soma-pds (T.s.m.) ; in the 
Atharvan, prathama-jd's and carshani-prd's (N.s.m.), and soma-pa (cf. Manu 
xi. 149). In the Rik, we have carma-mnd's (N.p.m. — can also be referred to B) ; 
in the VS., carma-mn&'m: in the Rik, reto-dhd's vrshabh&s; in the VS., garbha- 
dhd'm. 

Moreover, if we look at the later literature, we see that stems in radical a 
generally take the place of those in radical d. In the Veda we have the mascu- 
line stem suprajd'; but in Sanskrit, suprajd. In contrast with the Vedic mascu- 
line A-stems sdma-gd', agra-jd', ab-jd', giri-jd', rta-jnd', vdso-dd', pacu-pd', antari- 
ksha-prd', pathi-shthd', we find in the later language the B-forms : sdma-gasya, 
Ait. Br. ii.22; agra-jdm, R. ; ab-jeshu, Manu ; giri-jdya, Ait. Br. vii.l end; dharma- 
jMm, Manu; vdso-das (see Manu i?.229-232 for 16 compounds of -dd); pacu- 
pdndm, Brhat Samhita, xvi.14; antariksha-prd (V.s.m.), T.&r. iv.7.5 ; pathi-shthds 
and -shthdndm, MBh. Everywhere, the B-stems gain the upper hand. Similarly 
stems in & supplant those in #, p. 403, II. Compare Delbriick, Verbim, p. 87. 

By the aid of the lists of A and B-stems and of the enumerations of A and 
B-forms in the following pages, a more detailed comparison of the relative age of 
the A and B-forms might be made. 

On the other hand, we cannot doubt the existence of roots in 
short d. The facts of the related languages speak in favor of this 
view, and so do d priori considerations. Schleicher has asserted 
the originality of roots in short d (JSeitrdge, ii.92-99) ; and Del- 
brtick has given a most interesting discussion of the subject in 
his Verbum., pp. 87, 88, and 94, which see. Compare also 
Kuhn, Beitrtige, iv.202. 

G-r. puts stems like ddnu-dti' '(" ddnud-d") in the same category with vSd-a, cSd-a, 
Wb. 1698. I am uncertain whether this fact is to be interpreted as a tacit denial 
of Delbriick's view or not 

Now since the stems in radical short & are declined in all 
respects like stems in suffixal short 4, there arises a difficult ques- 
tion : Are we to assume that stems ending in radical short & were 
originally declined like those ending in suffixal short d, or not ? 
Is such a form as go-p&'m to be regarded as a direct relic which 
the Vedic language has inherited from the time of Indo-European 
unity, or are this and similar B-forms to be regarded as instances 
of a wide-spread transition from the declension of the stems in 
radical long d to that of stems in suffixal short d — that is, as tran- 
sitions from AtoB? 

In regard to some verbal forms with short d {dkhydt, dhvdt), 
Delbriick (p. 89) takes the view that they are direct relics of the 
pre- Vedic time (" dass die Kurze uralt uberliefert sei"). 



436 C. B. Lawman, [Had. & and 

In regard to the numerous forms of nominal compounds with 
short & (go-p&'m, etc.), the fact that they are, in general, demon- 
strably younger than those of stems in radical long d seems to me 
to point to a different theory. The existence of roots in short & 
is to be asserted for the primitive Indo-European language ; hut 
the existence of roots in short & is not to be admitted — except, 
perhaps, to a very small extent — for the oldest period of the Vedic 
language. 

When I say that roots " existed" at such a period in such a form, I mean, of 
course, Bimply that they were used for, or entered into verb and noun-formationB 
at that time in the one form or the other — for example, as -pd or as -pd. 

At the beginning of the time of separate Indian development, 
such roots were already for the most part out of use. Upon 
specifically Indian ground, however, the roots which in the oldest 
Vedic period ended in long d were gradually replaced by roots 
in short d. That is, I consider the Vedic root-forms with short & 
not as direct relics coming down from the time of Indo-European 
unity, but rather as new formations specifically Indian, which 
indeed coincide with the proethnic forms, but are not historically 
identical with them. They may therefore be properly designated 
as forms of transition from A to B, dating from the period of 
separate Indian development. 

The entire identity of form in certain cases of both series did 
much to increase the confusion of the two series (compare p. 370 
top), and thus to give eventually the upper hand to one or the 
other mode of declension, A or B, — here, naturally, to B, since 
its forms are vastly more frequent than those of A. This process 
of transition was the more easy, since it is in some of the most 
frequently occurring cases — the N.p.m., A.s.f., N.p.f. — that the 
forms of A and B are coincident. 

For the oldest Vedic period I would therefore assume that most of the radical 
stems were A-stems like soma-pd'. Doubtless all the declensional forms of such 
radical stemB belonged to series A. The N.p.m. was somapd's; and the N.s.m., 
without difference, somapd's. In the great majority of instances, however, a 
N.p.m. in -ds (dead's) corresponds to a N.s.m. in -ds (devd's); and after this 
analogy, the N. singular m. of the infrequent radical stems was also formed with 
short vowel (somapd's), to correspond to the N. plural m. (somapd's), and the forms 
of the two cases, N.p.m. and N.B.m., before alike, were thus differentiated. To 
the apprehension of the language-users, the stem became somapd', and upon thiB 
a system of B-forms was developed accordingly. 

So in the feminine, the A.s. crad-dhd'm answers properly to a N.s. prad-dhd's 
(a form of which the written text shows traces at vii.32.14 — see below, N.s.1) ; 
but since the vast majority of accusatives in -dm (citrd'm) answer to nominatives 
in -d (citrd'), the N.s. frad-dM' was formed after this analogy. In like manner 
we have the A.s.f. sva-jd'm, and correspondingly the N.s.f. divi-jd's; but also the 
N.s.f. sana-jd'. So in the plural we have the N.f. vdja-dd's; but, after the analogy 
of series B, the A.f. apva-dd's — not -dd's. 

To assume the existence of roots in short a simply on the basis of these com- 
pound noun-forms, when they can just as well be explained as instances of purely 
declensional transition to the fi-declension, would be a petitio principii. We must 
therefore seek for evidence of the existence of roots in & for the Vedic period on 
the field of the verbal flexion and elsewhere. 



a-stems.] Noun- Inflection in the Veda. 437 

The forms gatd, hath&s, etc. were supposed to give evidence of root-forms gS 
and ha; but Delbriick (p. 93) has deprived this evidence of its force by explaining 
the forms as a purely phonetic outgrowth of *gan-td, *han-thds, etc. (See also 
Brugman, Kuhn's Zeitsch. xxiv.257.) In like manner, forms like dkdyati and 
ddyate have been used to support the roots dhd, 'suck,' and di, 'mete out.' Even 
here, Delbriick (p. 165) comes to a different conclusion respecting the first, but 
admits the root dd. Similarly the a of the reduplicating presents, pibdsi, tishthdsi, 
etc., which Fick adduces as evidence of the original short a, is shown by Del- 
briick (p. 105) to be due to analogy (tuddmi : tuddsi : : pibdmi : pibdsi). Again, 
although pd'tis unquestionably implies a root pa, yet it is an entirely pre-Vedic 
formation and proves nothing for the Vedic period. 

After leaving out all the indecisive forms, there will probably remain some 
which forbid the complete exclusion of roots in a, from the field of truly "Vedic 
formations. Thus in viii.81.1rt, we have p&antam &' vo andhasas: why should 
we regard the aa as a mere phonetic resolution of d t is not the word perhaps to 
be divided pa-ant-am f 

It will require a careful, critical, and detailed investigation to determine how 
many of the forms apparently involving a root in U really imply a root in a. As 
a result of this determination we might be able to separate the roots for which we 
may assume a short a in the Vedic period (as dd, pa) from those for which we may 
not (as jd), and to call the compound noun-forms from the latter (as, a-jd's), 
" transition B-forms," and those from the former (as, madhu-pd's), " radical 
B-forms." This separation, however, is not a part of my task, and I have there- 
fore spoken of all the B-forms indiscriminately as coming from stems in "radical 
short d," and have enumerated them together in each case under "B." 

Series A is applied to masculines and feminines without distinc- 
tion, and in two or three sporadic instances to neuters. The 
grammars, however, state that the neuters of stems in radical d 
go according to series B n., and this is almost invariably the fact. 

The Masculine Stems. 

In the masculine, the forms of A and B are distinguishable for 
every case except the N.V.p., and the N.A.V. and I.D.Ab. d. In 
these cases the forms might be referred either to A-stems or to 
B-stems: thus ratna-dhd's (N.p.) may be assigned to the A-stem 
ratnadhd', or to the B-stem ratnadhd' (iv.34.8). Such doubtful 
forms I have referred to A-stems, unless the other unambiguous 
forms were referable only to B. Compare the articles pravdtejd, 
madhudhd, raghujd, rayidd, vayodhd', vdyugopa, svajd, in Grass- 
mann's Wb. Gr. says that stems in d may form the Lp.m. in 
-dbhis or -ais. I think he is wrong; see Lp.m. 

The general enumeration follows, and includes also the stems 
of many of the ambiguous feminine forms which may be referred 
to either A or B. Stems in brackets are from the Atbarvan. 

A. Stems of the A-forms. 

I. (a) The stem is identical with the root: gd' (gd, 'sing'); jd'; trd'; dd'; 
[dhd' ;] sthd'. 

(e) The stem is a compound of a radical substantive : parama-jyd' . 

II. (a) The stem is identical with a prepositional compound of the verbal root : 
adhi-pd'; paras-pd'; ni-shthd'; pari-shthd'. 

(c) The stem is a compound of a feminine substantive under II. (a) : aprajd'; 
bahuprajd'; suprajd'; Tcarmanishthd'; pwunishthd' (also -a') ; [dndvayd]. 

III. (a) The stem is a compound of a verbal root with a noun or adverb. The 
noun is most often a substantive in an accusative or locative relation (go-pd', 
ab-jd', apsu-jd') \ but it may be an adjective or adverb in some other adverbial 
relation (nava-jd", m-ahd'). 

VOL. x. 59 



438 C. B. Lanman, [Rad. d and 

(b) The stem is a compound of a word under III. (as) used substantively (devd- 



The ensuing list embraces the compounds falling under III. (a) and (b) ; in it, 
an affixed B signifies that the stem occurs also with d, and the hyphen is used to 
avoid repetition of the verbal. We have the following compounds : — with the 
verbal 

-krd' (kar, ' do') : d-krd — Gr., d-krd. 

-k/rd': dadhi- ; rudhi-. 

■kshd' (at. B): rbhu-1. 

■khd' (cf . B) : bisa- ; cf . khd, n. 

-gd' (gd, ' go'^ — cf . B) : adhva-, B ? ; and- ; tamo- ; tavd- ; pwo- ; samana- ; 
svasti- ; ogre- (Acv. Qr., see A.s.m.), B. 

-gvd (cf . B) : tia-gvd t. 

-gd' (gd, ' sing' — cf . B) : sdma-, B. 

■jd' (cf . B) : agra- ; adri- ; apsu- ; ab- ; udanya-, B 1 ; rta- ; rte- ; giri- ; go- ; 
tapo- f, B ; divi- ; divo- ; deva- ; dvi-, B ; nabho- ; nava- ; purd- ; p&rva-, B ; prath- 
ama-, B ; pravdte-, B ? ; bhareshu- ; manushya- ; raghu-, B ? ; vane- ; sana-, B ; 
sano-; saha-; saho- ; sva-, B?; svayam- ; [agni-; abhra- ; oshadhi- ; khala- ; 
nakshatra- ; vara-; vdta-; pakadh&ma-; Mranya-~\. 

-jnd': rta-; pada-. 

-trd': dn-agnitrd ; puru-trd ; ci p. 441 top. 

■dd' (dd, ' give' — cf . B) : anapvadd ; andfirdd ; abhiksha- ; agva- ; dtrna- ; ojo- ; 
go- ; jani- ; dravino-, B : dhana- ; bala- ; bhAri- ; rabho- ; rayi-, B ? ; vasu- ; 
vastra- ; vdja- ; vdso- ; sahasra- ; saho- ; svasti- ; havir- ; hiranya- ; [fata- ;] from 
"VS., prdna-, B; apdna-; vydna-; vwrco-; varivo-. 

-dhd' (dhd, ' put' — cf . B) : kiye- ; jani- ; dhAma- ; dhiyam- ; ndma-, B ; madhu- 1; 
ratna-,~B; reto-; vayo-; varivo-; vipo- ; sa/rva-; [adoma-, B?; yakshmo-; varco-\ 

■dhd' (dhd, ' suck') : payo-. 

-pd' (pd, ' oversee '—cf. B): abhipasti- ; rta-; kula- ; go-, B; (igopd, dhigopd, 
indragopd, devagopd, vdy&gopd, somagopd, sugopd' ;) chardish- ; jagat-; tanA- ; 
nishikta- ; papu-; vrata-; fevadhi-; sti- ; [nidhi-; sva-]. 

-pd' (pd, 'drink' — cf. B): ogre-; a/njas-; rtu-; (anrtvpd;) rd&- ; ktldla-; tapush- ; 
payas- ; p&rva- ; inadhu- (cf. madhvpd'tarna), B ; cvkrap&ta- ; pud- ; (rta-, B ; 
suta- ; soma-, B ; hari- ; havish-. 

-prd' (cf. B) : antariksha-, B ; kakshia- ; krshti- ; kratu- ; carshani-, B ; jarani- ; 
ratha-; rodasi-. 

-mld't: carmasmnd', B. 

-yd': rna- ; eva- ; jma- ; tura- ; deva- ; pubham- ; [andva-]. 

-rd': fata-rd't (text (atdrd). 

-sd' (san, ' win' — cf . B) : ap- ; acva- ; urvard- ; kshetrd- ; dhana-, B ; priya- 1, B ; 
vdja-; fata-; sadd-; sahasra-; wru-shd'; go-, B; nr- ; pacv,-, B; su- ; suar-. 

-sthd' (ci. B): rocama-; [rta-;] adhvare-shthd'; giri-; gharmye- ; nare- ; pathe- ; 
parvate- ; prthivi- ; barhih- ; mahhane- ; rathe-, B ; vakshane- ; vandane- ; van- 
dhure- ; hari-; hwrmye-; [ange- ; paihi-; bhuvane-; madhyame- ; rayi-; savya-\ 

-snd': ghrta-. ' 

B. Stems of the B-forms. Compare the neuter stems, p. 441. We 
have the following compounds : — with the verbal (or substantive) 

-kshd': dyu-kshli'. 

•kh&': su-khd'. 

-gd' (gd, 'go'): ogre-; adhva- 1 ; patam- ; sam- ; (ratha-samga ;) su- ; suar-; 
[a-sva- ; dcurh- ; rju- ; dur- ; vaXa- ; pitim-]. 

■gvd': atithi- ; eta-?; doc a-, ndva- (accent I) ; puro-gavd. 

-gd': (gd, ' sing') : sdma-gd, AT. 

-grd' (gar, 'swallow'): tuvi-grd. 

■jd' : a- ; udanya- 1 ; eka- ; tapo- ; p&rva- ; pravdte- 1 ; raghu- f ; sdkam- ; 
Uardytir- ; tarafl- ; tapo-; dvi-; ni- ; pratltama- ; samudra- ; stamba-; sva-]. For 
adhri-ja, see Grr. 

-id: d'-idf. 

-dd' (dd, 'give'): andnu-; ddnu- ; dravino-; rayi-t ; [ddyd- ; prdna-]. 

-da' (dd, ' bind') : rpya-. 



a-stems.] Noun- Inflection in the Veda. 439 

■dhti,': ratna- ; [adoma- 1 ; ndma- ;] garbha-. VS. 

■pd' (pd. 'oversee'): go-; mitho-avadyapa. 

-pd' (pd, 'drink'): madhu- ; ffta- ; [soma-]. 

-prS,': [dkUti-; kdma-; carshani-; prthivt-;] antariksha-, T.4r. 

-ml&'t: carma-mnd', VS. xxx.15. 

•88,' : dhanar ; priya- ; papu-sha ; go-sh&'-tama. 

-sihft.' : apnah- ; sarh- ; go-shthd ; tri- ; rathe-. 

-hti.' (han, ' slay') : ardti-, patru-, and sahasra-, in the AV. Of. neuters, p. 441 top. 

According to the grammars, adjective compounds of feminine 
substantives in & form the masculine stems in d: thus, an-avasd' 
(-ds yd'mas), from dvasd; a-fraddhd' (-d'npani'n) ; puru-nishthd' 
\-ds kaois, v.1.6); cf. dshthd. In the Veda, however, this rule is 
by no means general : thus, we have the nominatives s.m. bahu- 
prajd's, su-prafd's, puru-nishthd's (viii.2.9) ; and the A. s.m. karma- 
nishthd'm. This wavering between A and B corresponds to that 
of the feminines between the A and the B-forms. In both m. and 
f. the B-forms prevail in the later language {labdha-nidrds, 
Kathits. iv.9). 

The Feminine Stems. 

In the feminine, the forms of A and B are not distinguished 
from each other in the A. and Lb., in the I.D.Ab.d., and in the 
N.,L, D.Ab., L., and V.p. — that is, in some of the cases of most 
frequent occurrence (they are marked with an asterisk in the table 
on p. 434). The N.s. and A. p. are distinguishable ; and in the 
N.s. we have mostly A-forms (with -s), and in the A.p., exclusively 
B-forms (with -ds — not -ds). In point of fact, therefore, there are 
no feminine forms belonging distinctively to A, except in the N.s. 
In this case, the A-forms are about equal to the B-forms in the 
Rik, but the B-forms belong to late hymns ; in the Atbarvan, the 
B-forms decidedly outnumber the A-forms in the N.s., as well 
as throughout the rest of the declension. 

The declension of masculine radical Osteins like ddnu-dd' is the 
same as that of B m. (ga-td') ; but since the feminine of such a 
stem would be ddnu-dd', it is plain that this may be declined 
according either to A, or to B f. And so we have the same 
wavering as in the masculine : thus, gopd's (N.s.m.) is to gopd's 
(N.s.m.) as devdgopds (N.s.f.) is to devdgopd (N.s.f.). It seems 
that the radical (A-) declension went out of use for feminines at 
an early period. 

Theoretically, all the compounds enumerated under division III. 
of the masculines ought to be declined in the same manner in the 
feminine. According to the grammars, the N.d.f. should be like 
the N.d.m. (purd-jd') ; but in reality we have here B-forms (rdOirp'e, 
pilrva-jS). In discussing the masculines, we found it probable that 
most of the forms like gopd's were to be regarded as instances of 
transition to B m. Here, also, it seems easier to assume a general 
and wide-spread system of transitions to B f. Otherwise, we must 
assume that every fem. stem (e. g. devdgopd-; rdtipd'-; pdrvajd'-) 
to which we refer the B-forms (e. g. devdgopd, N.s.f. ; rdiXp'e, 
pilrvaj'e, N.Ad.f.) is feminine to a masculine stem with short d 



440 0. R. Lanman, [Ead. d and 

(e. g. devdgopd-, rdttpd,', pttrvqjd') ; this is possible (compare the 
articles sanajd, -jd! ; dnagnitrd; devdgopa, -pd; ndvagva; sahd- 
gopa; sugd; svdgopd; svajd; svdsaslhd) ; but does not seem 
generally warrantable. The masculine stems are themselves tran- 
sition-stems, and the feminine forms can just as well be considered 
as direct forms of transition from A to B. 

A. The A-forms. Using the same categories as above, the only 
feminine forms belonging distinctively and unequivocally to A 
are as follows : 

I. kshd's, gnd's, and vrd's (N.s.i) ; kshe f (D.s.f.) ; gma's and jm&'s (Ab.G.8.f.). 
III. riaipd's ; rtejd's ; godhd's ; goshd's ; divijd's, divojd's ; vanejd's ; patasd's ; 
sahasrasd's; samanagd's; [kiUapd's, gopd's; vayodhd's ; vasurfd's], — all N. S.I. 

B. Stems of the B-forms and of some of the forms that may be 
referred either to A-stems or to B-stems. 

I. (a) lcshd'; kshmd'; hhd'; grid'; jd'; jyd', 'bowstring;' rd'; vrd'; dhyd'. 

(b) The stem shows a reduplicated form of the root : gdngd ; janghd. 

II. (a) Almost all the stems of this category are substantives and none shows a 
form belonging distinctively to A. Gr. derives prajd', s.v., from jam with pro by 
the " formative suffix d." It is declined indeed as if the a were suffixal ; but this 
mode of declension may be secondary, and I assume an original root-stem in -jd. 
He says, Wb. 675, " In ni-dhd', go-dhd', prad-dM', sva-dhd', the suffix d has been 
added, as the declension shows." But we shall see that with some of these words 
there are even yet traces of the A-declenskm. Enumeration: dbhikhyd'; \wpagd', 
dwgd' ;] prajd', [upajd', parijd']; [prajM' ;] d'-td (f tarn, — accent I); nidrd' ; 
apadhd', durdM', nidhd', praddhd', sva-dhd' — root dhd, ' put' — , [godhd', tirodhd', 
purodhd'] ; pratidhd', svadhd' — for *su-dhd', dhd, ' suck ;' prapd' ; abhibhd', 
te&bhd f, vibhd', sabhd'; praiimd', pramd'; dpayd'; [wpavd', pravd';~\ avasd ; ava- 
sthd', parishthd', pratishthd', vishthd'; prahd'. Several are used as adjectives : 
anushthd', nishthd', parishthd'. 

(c) Compound : asarhjrid, AV. 

The Nbutee Stems. 

Aside from a few sporadic words (p. 445), the stems of the 
neuter forms always end in short d, and the forms coincide with 
those of series B n., page 434. 

Declensionally, then, they might have been treated with the 
neuter <3S-stems (pp. 331-354) ; and a few have been there included : 
thus, antdriksham has been counted on p. 331; and similarly 
trishadhasthd-s, N.s.m., on p. 329 ; and trishadhdsthd, N.s.f., on 
p. 355. 

This inconsistency is the less reprehensible inasmuch as the final 
root-element of a compound often loses its distinct character as a 
root : thus, in pr-shthd! we have as final element the root sthd or 
sthd; but in the compounds vitdprshth&s, -prshthd, ghrtdprshtham 
(barhis), the final syllable can hardly have been felt as a radical. 
Quite analogously the recessive accent of the English compound 
fore-head (pronounced fb'r-M) has, to the apprehension of the 
vulgar, entirely obscured the fact that the word has anything to 
do with head. In some words, as updstha, the character of the 
final element is uncertain. 

A complete list of the neuter stems whose final element is a simple radical 
belongs rather to the chapter on the formation of steins (p. 325 — see Lindner, p. 
26). These remarks apply also to the radical masculines in & and their feminizes 



a-stems.] Noun- Inflection in the Veda. 441 

in d (of. articles sana/jd', sv-dsa-sthd'), so far as they have the forms of series B ro., 
or of B f. Thus vdydgopds has been counted among the 1954 forms mentioned on 
p. 344 end ; although it ought not to be separated from papupd's, etc , N.p.m. I 
do not think that many words properly to be classed here have been left out. 

The enumeration follows : 

Dyu-kshd (antdri-ksha) ; tuvi-kshd ; kha ; dw-gd, su-gd ; [asthi-jd, tanA-jd, pro- 
thama-jd ;] rtd-jya; dnsa-tra (cf. lanu-tra, and see G-arbe, Kuhn's Zeitsch. xxiii. 
476,480); [adoma-dd;] ararin-da (dd, 'bind'); [drat, uttara-drd, krshna-drd ;] 
sva-dhd (anushvadhdm) ; [vdtdgopa, paras-pa (VS. xxxviii.19);] madhu-pd; kdma- 
prd ; su-mnd ; pr-shthd, bhayd-stha, mahd'-vailastha, sadhd-stha ; karanja-hd, 
vrtra-hd, satrd-hd. 

Stems whose formation is exceptional or doubtful. A few 
such remain. We have from the adverb tdthd the stem d-tathd — 
N.s.m., dtathds, i.82.1. The Veda has the forms: d'tds, N.p. ; 
d'tdsu; and d'tais. Gr. refers them all to a m. stem d'td. Against 
the view of Gr. speaks the fact that all nouns substantive like 
prajd' are feminine. His dictum that series A admits both end- 
ings, -dbhis and -ais, is supported only by dhanasais. Perhaps, 
then, we ought, with BR., to refer d'tais to d'td, m., and the other 
forms to d'td, f. The f. stem maryd'dd is unclear. 

It does not help us in the least to refer the N.s.m. updnd, with 
Gr., to a stem updnd, since a form without s is no more admissible 
here than in the N.s.m. of the stems in -as, which see. 

The strong forms of the word for ' path' are as follows : 

in BV., pdnthds, 31 in AV., pdnthds, 3 panthd, 1 

pdnthdm, 31 pdmthdm, 19 pdnthdnam, 1 

pdnthds, 6 pdnthdnas, 1. 

The N.s.m. dtHrthapanthds occurs twice, and purupdnthds 
(N.s.m.), once. In like manner we have mdnthdm, never mdnth- 
dnam. Once, in a hymn of by no means antique stamp, i.100.3, 
we have the form pdnthdsas. The Zend forms correspond in the 
main to those of the Rik : 

[*pafitdo] pailtS., 1 

parttdrn ) ,, paritdnem, 1 (Vend.) 

pathdm j pantdnd, 2. 

In i.127.6 g and h : v.10.1 and viii.57.13, Kuhn and Gr. propose to read pam- 
thaam, and in viii.31.13, pdntha&s. In each case, the word is at the end of a 
catalectic pdda of 7 syllables and the resolution is unnecessary. If, however, any 
one insists upon having 8 syllables, he must read, e. g., urum rdthdya pdnthdnam 
(pdnthdnas), and in no case pdnthdnam (pdnthdnas). See Kuhn, Beibrage, iv.203. 

In like manner the two resolutions in the Avesta (Geldner, Metrik, p. 17) are 
unnecessary, since they are at the end of catalectic pddas of 7 syllables (Yt. 
viii.7: x.86.). 

The Eik, then, does not have a single form that gives direct evidence of a stem 
pdnthan. Not until we get to the Atharvan do we see any forms involving this 
stem unquestionably : N.p., pdnthdnas, to the exclusion of pdnthds ; A.s., pdn- 
thdnam, once, against pdnthdm, 19 times ; and N.s., panthd, iv.2.3 (where pada 
reads panthd), against pdnthds, thrice. Similarly in the Avesta, pafltdnem occurs 
only once, and that in the Vd. ; pantdnd, twice ; and pafitd (i. e. pantd, from a stem 
pantan ; stem pafitd would form pafltdo), Vd. xvi.2. The evidence, both from the 
Veda and the Avesta, pronounces the forms from the stem pdnthan the younger. 

Brugman, however, explains pdnthdm as arising by contraction from panthdn-u 
(Studien, ix.307). Prom pdnthdm, the language-users got by "false abstraction" 



442 C. R. Lanman, [Bad. & and 

the stem pantM-, and so the N.s. panihd-s was formed, and the N.p., panfh&s, i. e. 
*pcmthd-as. The only analogies for such contractions with ra-stems are rnanthdm 
and the somewhat doubtful mdhd'm (*mahan-li). If this is the correct explana- 
tion, it would seem as if the stem panthan ought to have left some direct traces in 
the Rik. The a of parttam does not necessarily imply a form paSUom-v. (see 
Schleicher, Cornp. 4 §21.7). If the N.s.m. pdnthds is an Associations-bildung based 
entirely on the A.s.m., we must suppose that no nominative could have been 
formed until after the form of the accusative was settled. 

On the other hand, if we assume out and out a stem panfh-d (maniK-d), we 
cannot regard the a as a derivative suffix, for that is unparalleled. 

Possibly the peculiar treatment of the two stems p&nthan and 
mdnthan is to be referred to the analogy of the root-words ending 
in -an below, which exhibit parallel forms in -d. I have therefore 
put the strong declensional forms of these stems in this chapter, 
assuming transition-stems pdnthd, mdnthd, and making pdnthd 
related to pdnthan as -jd! to Jan. Somewhat similar transitions 
are seen in an-asth&'-s (cf. an-asthdn — N.s.m., -d') and an-akshd'sas 
(cf. akshdn) ; see ara-stems, N.A.s.n. 

Without interpreting the facts, I would observe that four 
verbals correspond to roots with nasals : -khd', khan; -gd', gam; 
-jd', jan ; -sd', san. See Delbrfick, pp. 92, 93. The relation is 
perhaps similar in -gvd and gvam, d'-td and tan, and the words 
for ' earth,' kshd' and kshdrn. Cf. further, gmd ', gam; jmd',jdm; 
kshmd', kshdm; dhmd, dham; mnd isu-mnd,'), man; vrd (vrd'), 
war; kvd (d-krd), kar; gra (tuvi-grd }, gar; trd, tar; and see 
Curtius's Studien,v.l97. 

Several roots appear even in a triple form. We have : go-shan-as 
(napdt), go-shd'-m (sSmam), go-shd'-tamds; rbhu-kshdn-am, rbhu- 
ksh&'-s, dyu-kshti'-s (cf. kshd-trd, kshi) ; tdn (tdnd gird'), d'-td, 
d'-td. To these we may add the root han. In the Veda occurs 
vrtra-hdn-am; the Zend vhrUhra-jdo, Yacna lvii.14 etc., corre- 
sponding to Skt. *vrtra-hd'-s, supports the second form hd; while 
the Atharvan words patru-ha'-s, ardti-hd'm, etc., supply the third. 
(The form -ha is again brought back to the consonant declension 
by an added t in samhdt.) The Zend -jdo may, however, repre- 
sent *-jan-s (= -hdn-s, -hd') ; cf. zdo, *zam-s — G.s., z(e)m-o— , and 
zydo, *zyam.-s. Important, finally, from an exegetical point of 
view is Grassmann's suggestion, Ueb. i.576 : he assumes a stem 
da'=ddm, and of this he takes de (v.41.1) as L.s.n., which see. 

The following synopses exhibit the terminations of the 
inflectional forms : 

Series A. Masculine. 

Singular: N., -ds, -aas; A., -dm, -aam; D., -e; G., -as; V., -ds. 
Dual : N.A.V., -d, -au. 

Plural: N.V., -ds, -aas, -dsas ? ; A., -dsf, -ds?? ; I., -dbhis, 
-ais??; Ab., -dbhyas; L., -dsu. 

Feminine. 
Singular : N., -ds, -<f ; A., -dm, -d'; I., -d; D., -e ?; Ab., -as; 
G., -as. 

Plural: N.V., -ds; I., -dbhis; D., -dbhyas; L., -dsu. 



a-stems.] Noun- Inflection in the Veda. 443 

Nbutbe. 
Singular : N.A., -ds; D., -e, -ai (inf.). 

The terminations of series B m. and n., and of series B f., 
coincide — so far as they occur at all — with those given on page 
329 and on page 355 respectively. 

Dissyllabic forms of monosyllabic stems do not shift the 
accent to the ending. Only a few occur : kshd'-su ; jd'-su ; 
gnd'-bhis, gnd'-su ? ; jydyd, jyd'yds ; dhydyd ; kshmayd' ( ! an 
isolated exception). Compare p. 421. 

Nominative Singtolak Masculine. 

A. The ending -s is added to the stem : as, adri-jd'-s. The 
form occurs in the Rik 273 times (from 102 stems) ; and in the 
Atharvan, 55 times (from 26 stems). In this and the following 
lists, Atharvan forms found only in verses quoted from the Rik 
have not been included. Enumeration : 

I. gd's, 'singer,' x.127.8 (? of. AV. xii.].39); jd's, ix.93.1(cf..;'<2's-pafo's, vii.38.6); 
dd's, vi.16.26 ; sthd's, x.35.9 ; dhd's, TS. ii.6.4 4 ; — parama-jyd's, viii.79.1. 

II. adhi-pd's, 2 (AV., 8) ; paras-pd's, 3 ; ni-shthd's ; — bahu-prajd's; su-prajd's ; 

— puru-nishthd's ; — dUre-bhds (? see rad. *-stems, N.s.m.); — an-dvayds, AV. 
vii.90.3. 

III. (arrangement by roots) dadhi-hrd's, 4 ; — rbhu-kshd's, 14 ; — bisa- 
khd's ; — and-gd's ; puro-, 3 ; samana- ; — sdma-gd's ; — adri-jd's (see 
iv.40.5) ; apsu- ; ab- ; rta-, 2 ; rte-, 2 ; go- ; divi- ; deva- ; dvi- ; nabho- ; purd- ; 
p&rvctr ; prathama-, 4; manmhya- • vane-; sana-; sano- ; salw,- (x.84.6) ; sake-, 
2 ; — rta-jfid's, 3 ; — dtma-dd's ; ojo- ; go-, 4 ; draviifo-, 9 ; dhana-, 3 ; bala-, 
2; bh&ri- (iv.32.19,21); sahasra-; saho-, 3; suasti-, 3; hiranya- ; — hiye-dhd's, 
2 ; dhdma- ; dhiyam-, 2 ; ndma- ; ratna-, 8 ; reto-, 3 ; vayo-, 1 2 ; sarva- ; — abhi- 
fasti-pd's, 2 ; rta-, 2 ; go-, 33 ; a-gopds ; devd-gopds ; aoma-gopds, 2 ; sugopd's, 2 ; 
tanti-pd's, 5 ; pacu-, 4 ; vraia-, 5 ; fevadhi- ; sti- ; — r tu-pd's, 3 ; dn-r tupds ; 
p!),na-pd's, 2 ; suta-, 5 ; soma-, 5 ; hari- ; — antariksha-prd's, 2 ; hratu-, 2 ; 
carshani-, 1 • rodasi- ; — rna-yd's, 5 ; eva- ; tura- ; deva-, 4 ; — ap-sd's, 3 ; 
acva-, 2; dhana-; vdja-, 2; fata-, 3; sahasra-, 6; go-shd's, 2; nr- ; su- ; suar-, 
6; — rocana-sthd's ; adhvare-shthd's ; giri-, 2, andiz.18.1; manhane- ; rathe-, 2; 
vandane- ; vandhvire- ; hari-, 6. 

Gr. proposes: giri-shthaas, ix.62.4; 98.9; ghrta-snaas, viii.46.28 ; dravino-ddas, 
vii.16.11 : viii.39.6. These resolutions are all at the end of pddas of 1 syllables 
and needless ; read -ds. 

We have from the Atharvan 46 forms (from 24 stems) : dbhra-jd's, i.12.3 ; 
deva-, x.6.31 ; nakshatra-, vi.110.3; prathama-, iv.35.1: vi.122.1: xii.1.61 ; vdta-, 
i.12.3; hiraipya-, iv. 10.1,4; — dravino-dd's, xix.3.2 : xx.2 4; dhana-, iii.15.1 ; 
havvr-, v.1.7; — reto-dhd's, v.25.1 ; vayo-, v. 11.11 : vii.41.2: ix.4.22: xii.3.14: 
xiii.2.33: xviii.4.38: xix.46.6; varco-, ii. 1 1.4; — abhifasti-pd's, ii.]3.3: iv.39.9: 
v.18.6: viii.7.14: xix.24.5,6; go-, 4 times: as, iii.8.4; tanfa-, vi.53.2 ; nidhi-, 
xii.3.34,41,42 ; sva-t, iii.3.1; — soma-pd's, viii.5.22 ; — dhana-sd's, xir.31.8; 

— rta-sthd's, iv.1.4 ; ange-shthd's, vi.14.1; bhuvane-, ii.1.4: iv.1.2 ; madhyame-, 
ii.6.4 : iii.8.2 ; savya-, viii.8.23. The VS., xvii.15, has prdna-dd's, apdna-, vydna-, 
varco-, varivo-. For others, see iii.16. 

Resolution is necessary in RV. x.64.10, rbhu-kshdas; so AV. 
xix.55.3,4, vasu-ddas; likewise in RV. iv.6.3c, 4c, nava-jdas and 
papu-pdas (Gr., -d'as). 

In no wise is a IST.m. form in -d admissible. In ii.20.2, the 
metre shows that we must pronounce abhishtipd 'asi jdndn. If 



444 C. R. Lanman, [Rad. & and 

we take abhishtipd! as a compound, we shall have to regard the 
samhitd-reading as an instance of elision and crasis. Perhaps it 
is better to read dbfmhti pdasi, ' thou protectest with thy favor.' 
See p. 381 top, and Gr., s.v. In v.87.1-9, evayd'marut, Gr. sug- 
gests that evayd' may be a N.s.m. ; but here we must enter a non 
liquet. For dtathds and pdnthds, see p. 441. 

B. The Rik has 24 forms (from 14 stems) ; the Atharvan, 29-f- 
forms (from 24 stems). Enumeration : 

Dyu-kshds, 4 ; — su-Jchds; — agre-gds; patam-; su-, 4 ; — ddpa- 
gvas f; ndva-, 2 ; puro-gavds; — a-jds, 2 ; adhri-; — andnu-dds, 
3 ; ddnu-; — apnah-sthds; purunishthd' s. Here belongs, prob- 
ably, papu-shds, v. 41.1 ; see G.s.m. 

From the Atharvan, dpum-gds, vi.14.3 ; rju-, i. 12. 1; patam-, 
vi.31.3 : xiii.2.31 ; 3.1; pitim-, xi.5.12 ; vala-gds, xii.5.39 (? Ind. 
Stud, iv.304) ; — eka-jds, x.8.5 bis; jardyu-, i.12.1 ; ni-, iii.5.2 ; 
prathama-, iv.4.5 ; samudra-, iv.10.4; stamba-, viii.6.5 ; — ddyd- 
dds, v.18.6,14 ; prdna-, iv.35.5 ; — drds, vrds, nyds, f xi. 7.3,4 ; — 
ndma-dh&'s, ii.1.3 (Rik x.82.3 has ndnia-dhd's) ; — dkilti-prds, 
iii.29.2 ; kdma-, xi.7.8 ; carshani-, iv.24.3 ; prthivi-, xiii.2.44 ; — 
go-shthds, iii.14.6, etc. ; — patru-hds, i.29.5 : vi.98.3 ; sahasra-, 
viii.8.'l7. 

Nominative Singular Feminine. 

A. The form is like that of the masculine. It occurs in the Rik 
18 times (from 13 stems) ; and in the Atharvan, 5 times (from 4 
stems). Enumeration : 

I. kshd's, 5 ; vrd's, i.124.8 (Roth and Say. as pi.) ; gnd's (see p. 
355). Cf. bhd's, BR. v.234. 

III. rta-pd's; rtejd's; godhd's, x.28.11; go-shd's,x.95.S; divi-jd's; 
divo-jd's; vane-jd's; pata-sd's,x.95.3; 178.3; samana-gd's; sahas- 
ra-sd's, x.178.3. 

From the Atharvan, kula-pd's, i.14.3 ; go-pd's, xii.1.57 ; vayo- 
dhd's,ix.l.8: xviii.4.50 ; vasu-dd's, xii. 1.44. 

(a) Although no other forms than the above show a s in the 
pada-pdtha, we do yet find the following words standing before 
a vowel with hiatus in the written text, and are warranted in 
assuming that these words originally formed their nominative 
with -s. They are: jyd', vi.75.3 ; ni-drd', viii.48.14; pra-pd', 
x.4.1 ; prad-dhd', vii.32.14 ; sva-dhd', x.129.5. To Ibese add 
prati-md', vi-md', un-md', TS. iv.4.11 s ; and see TPr. x.13, and Ind. 
Stud, xiii.104, note 2. 

It was doubtless the case-ending -s that kept tjiese words from combining with 
a following vowel ; but since the transition of the radical feminines to the B- 
declension became general at an early period, the full form was felt to be, e. g., 
pra-pd' rather than pra-pd's, and was so entered in the pada-text. In order to 
establish the apparently anomalous relation of the two pdthas in regard to these 
words, it became necessary for the Pr&t. to mention them expressly (ii.29). If 
we compare the two pdthas in regard to these nominatives, we find that the Prat, 
is complete and exhaustive. 

If, however, we compare the written saihhitd with the oral tradition or 
metrically spoken text, we find that the former is not consistent with itself. It 
has svadhd'sU at i.165.6, and svadhd'mitd at v.34.1, where we must pronounce 
svadhd' dstd and svadhd' amitd (cf. x.129.5<2). In like manner we saw at p. 356 



a-stems.] Noun- Inflection in the Veda. 445 

that the metre requires hiatus between the -d of the N.s.f. and a following vowel 
in 28 instances ; but that the diaskeuasts and after them Qaunaka have taken 
cognizance of this fact only in the case of tshd' and mantshd' (five instances). 

Since the Prat, cannot be convicted of incompleteness in enumerating anomalies 
like prapd', and since it does not mention godhd', x.28.11 (Muller's editions 1 and 
2), tibhyo godhd' aydtham karshad eiat, we must conclude with Aufrecht (Bigveda?, 
preface, p. v) that the true ^ada-reading is godhd'h. This is the A-form, and in 
the preceding verse we find with glaring inconsistency the B-form godhd' ; this 
also occurs at viii.58.9 and AV. iv.3.6 (nimruk te godhd' bhavatu, p. godhd', 
without avagraha and without visarga — ef. Ind. Stud, iv.305-6). 

B. The form is without -s, like the N. of stems in suffixal -d. 
It occurs 16 times (from 14 stems). 

Enumeration: abhi-bhd' ; kubhdf; godhd!, viii.58.9 : x.28.10 ; 
devdgopd, vii.35.13 (Mtlller, -pdh ; why is this form any less 
grammatical than the N.s.m. suprajd' s ?) : x.63.16 ; diu-kshd'; 
pra-jd'; prati-md'; pra-md' ; madhu-dhd' ; maryd'dd ? ; sana-jd' ; 
svdgopd; sva-dhd', ix.ll 3.10; sva-dhd', i. 176.2. 

We have from the Atharvan 27 forms (from 9 stems) : dpa-gd 
(BR. apagd'), i.34.5 : ii.30.1 : vi.8.1-3 ; dur-gd', xii.4.23 ; jd', 
v.11.10 bis; upa-jd', xi.1.19; d-saihjnd, xii.5.34 ; godhd', iv.3.6 ; 
prad-dhd', 13 times: as, v.7.5 ; devdgopd, vii.20.5 : xix.11.3; 
pra-pd', iii.30.6 ; from TS., md', iv.3.7 1 . 

Nominative and Accusative Singular Neuter. 

According to the grammars, stems in radical long d shorten 
that vowel in the neuter : thus, pankha-dhmd'-m. 

A. About eight forms occur with long d, which are used as 
neuters, and end in -s. Since a true neuter form with masculine 
case-ending is quite unexampled, we are perhaps forced to the 
conclusion that masculine forms have here been used in default of 
any others save those with short d. See p. 377 end. They are : 
pataysd's (vdcas), vii.8.6 ; suprayd's (barhis — cf. vi.63.3, suprdya- 
ndtamam), vii.39.2 ; sthd's, ii.27.4 : x.88.4 : (jdgac ca) i.80.14; 
indra-jd'h soma-jd' dtharvandm asi vydghrajdmbhanam, AV. 
iv.3.7; preshtha-sthd's, Tandya Br. vii.8.2 (BR.). 

Delbruck, Verbum, p. 228, mentions pratidhdm (sic) as an infinitive in -am from 
vocalic root. If he means pratidhd'm, AV. viii.8.20: xi.10.16, would it not be 
better to take it as ace. of a nomen actionis, on account of the form ? 

B. See page 440, and compare p. 407 top. Here belong 78 or 
more forms (from 9 or more stems) : khdm, 2 ; tuvi-kshdm; dur- 
gdm; dyu-Jcshdm, 4 ; (prshthdm, 10;) vrtra-hdm (pdvas), vi.48.21 
bis; satrdrhdm (pa-unsiam), v.35.4 ; sadhdstham, ? 14 times; 
su-gdm, 14 ; su-mndm, 39. 

From the Atharvan : adoma-ddm, vi.63.1 ; krshna-drdm, ix.7.4 ; 
prathama-jdm (ojas), i.35.2 (and VS. xxxiv.51) ; vd'tagopam, 
ii.12.1; su-gdm, xiv.2.6. 

Accusative Sinoular Masculine. 

A. Here belong 115 forms (from 59 stems). The AV. shows 9 
forms (from 7 stems). Enumeration : 
vol. x. 60 



446 C. R. Lawman, [Had. d, a-stems. 

I. jd'm, ix,89.2 ; trd'm, 2 times ; sthd'm. 

II. niskthd'm ; karma-nishthd'm, x.80.1 ; pari-shthd'm. 

III. dadhi-hrd'm, 8 ; rudhi- ; — tamo-gd'm ; tavd- ; suasti-, 2 ; — sdma- 
gd'm ; — agrayd'm ; ndbho- ; purd-, 2 ; prathama-, 2 ; bhareshu- ; — dn-afva- 
ddm ; an-df$r-ddm ; a-bhiksha-dd'm ; jani- ; dravino-, 1 ; dhana-, 2 ; rabho- ; 
vasu- ; saho-, 4 ; — vayo-dhd'm, 3 ; varivo- ; vipo- ; — abhi&asU-pd'm ; go-, 1 1 ; 
nishikta- ; vrata-, 2 ; — aHjas-pd'm ; prta- ; soma-, 3 ; — antariksha-prd'm ; 
carshani- ; ratha- ; rodasi- ; — ap-sd'm, 2 ; wrvard-, 2 ; dhana-, 2 ; sadd- ; 
sdhasra-, 6 ; go-shd'm ; smr-, 4 ; — rocwna-sthd'm ; giri-shthd'm, 4 ; nare- ; 
pathe- ; pcwvate- ; barhishthd'm ; rathe-shthd'm, 3 ; hari-. 

The form in -d'm stands at the end of a catalectic pdda, which Gr.— by the 
unnecessary resolution, -dam — makes acatalectic. At the end of pddas of 1 
syllables, we have : Tcdkshia-prd'm, yiii.3.22; tanQrpd'm, viii.60.13; paihe-shthd'm, 
v.50.3; ratha-prd'm, viii.63.10; ralTie-shthd'm, viii.33.14: — at the end of a pdda 
of 11 : giri-shthd'm, be. 85. 10. 

"We have from the Atharvan : praihama-jd'm, ii.1.4 ; — havir-dd'm, vii.78.2 ; — 
ratna-dhd'm, vii.14.1 ; — go-pd'm, xiii.2.2 : ix.10.11 (Eik i.164.31) ; — apa-vd'm t, 
xix.50.4; — pathi-shthd'm, xiv.2.6; rayi-shthd'm, vii.39.1 ; 40.2; from Acv. Qr. 
v.5, agre-gdm (Eik only agre-gas). 

Resolution is necessary in kshetrd-sdam, iv.38. 1 ; and in ab-j4'am, 
vii.34.16a (aksharapankti). 

For pdnthdm, mdnthdm, see p. 441 ; for mahd'm, see stems in 
an. 

B. The Rik has 26 forms (from 10 stems) ; the Atharvan, 8 forms 
(from 5 stems). Enumeration : atithi-gvdm, 6 ; andnu-ddm; 
eka-jdm; go-pdm; tri-shthdm; ddga-gvam; dyu-kshdm, 5 ; patarh- 
gdm, 2 ; madhu-pdm; su-khdm, 7 ; from the Atharvan : su-gdm, 
iii.3.5 : xiv.1.58 ; 2.8; d-sva -gam, xii.5.45 ; eka-jdm, vi.99.1 ; vala- 
gdm,v.Z\A: x.1.18; ardti-hdm, xix.35.2 ; so garbha-dh&'m, VS. 
xxiii.19 : TS. vii.4.19 1 ; carma-mnd'm, VS. xxx.15. 

Accusative Sinqtoar Feminine. 

The forms of series A and B are here coincident. The Rik has 
81 forms (from 27 stems). Enumeration : 

I. ksham, 13; khd'm, 2; gnd'm; jd'm, x.3.2; jyd'm; rd'm 
(? p. 431) ; vrd'm, i.121.2 ;— jdnghdm, 2 ; 

II. dva-sdm ; pra-jd'm, 22 ; prati-shthd'm, 2 ; pra-hd'm ; sa- 
bhd'm, 2 ; — siirprajd'm; 

III. antariksha-prd'm; apva-sd'm; uru-shd'm; rta-jnd'm; dur- 
dhd'm; dhana-sd'm; vdja-sd'm; prad-dhd'm, 6 ; sva-jd'm; Isva- 
dhd'm, 11 ; 2sva-dhd'm, 3 ; suar-shd'm; — d-gopdm. 

"We have from the Atharvan 12 forms (from 8 stems) : kshd'm, 
v.i.5 ; tiro-dhd'm, viii.l0.286«'sy prati-dham (see p. 445), viii.8.20 : 
xi.10.16 ; yakshmo-dhd'm, ix.8.9 ; prad-dhd'm, x.2.19 ; 6.4 : xix. 
64.1 ;' upa-vd'm, pra-vd'm, xii.1.51 ; pra-hd'm, iv.38.3. Cf. bhd'm, 
£at. Br. ix.4.1°. 

The word kshd'm occurs furthermore in Rik i.67.5 ; 174.7 : 
vi.6.4 : x.31.9, and has the value of two syllables. Gr. proposes 
kshd'm-am; see stems in radical m, A.s.f. 

Elision and crasis. In viii.32.6, drd'd upa svadhd' gahi, Roth 
makes svadhd' =svadhd'm d'; Gr. sees an I.s.f. 



I.s.m.] Noun- Inflection in the Veda. 447 

Instrumental Singular Masculine. 

A. There is no example. 

B. Only example : rathe~$hthena,ii.l'7.3. 

Instrumental Singular Feminine. 

A. Here belong 10 forms (from 7 stems). They might also be 
referred to B as homophonous instrumental (p. 358). 

They are: apa-dhd, ii.12.3 (where the Abl. apa-dhds, written 
apadhd before valdsya, is needed — Gr.) ; abhi-khyd' , thrice ; 
a-sthd", as adv., x.48.10 ; jmd' ; prati-dhd', viii.66.4 ; svadhd', 
vi.2.8 : viii.32.6 (see A.s.f.) ; prati-shthd' , x.73.6. The VS. has 
dptr-dd' at viii.5 ; and the TS., at iii.2.8', the form dptr-ddyd'. 
Unless we assume a stem -ddyd' (?), and take this as a homo- 
phonous I.s., both the quantity of the penultimate d and the 
accent are anomalous. See BR., s.v. 

In sdkdm pratishthd' hr'did jaghantha, x.73.6, I take pro- as 
instr. of accompaniment : ' Along with that on which she rested 
(sc. dnasd ushdsas), thou smotest her in the heart' (hrdi d', BR.). 

The adv. suffix -dhd is taken by BR. as an instrumental. 

B. Here belong 52 forms (from 9 stems) : hdbhayd? ; ni-dhdyd, 
2 ; hshmaydl (accent !), 5 ; jydyd; dhydyd; pra-jdyd, 1 1 ; prad- 
dhdyd, 4 ; sva-dhdyd, 26 ; sva-dhdyd. 

Instrumental Singular Neuter. 

A. There is no example. 

B. Here belong : rtd-jyena ; kdma-pr'ena ; su-mnhtia; from 
AV., kdma-pr'ena, x.9.4 ; su-gena, xiv.2.11 (Rik x.85.32, sugibhis). 

Dative Singular Masculine. 

A. Here belong 9 forms (from 7 stems) : ktldla-pe, x.91.14 ; d'e, 
v.41.1 (but see L.s.m.) ; dhiyam-dhb, vii.13.1 ; papurshb, i.127.10 ; 
$wi-p'e, x.100.2 ; $ubham-ye, iv.3.6 ; havir-di,, i.153.3 : iv.3.7 : 
vii.68.6. For updne, see stems in an, D.s.m. 

The infinitive prad-dhe, i.102.2, is to be regarded as dative in 
form as well as in meaning. Since the meaning of ix. 70.4ft admits 
of a dative, it is better to regard pra-m& as a dative also in form. 
BR. take it as a locative form, and this is certainly possible ; for 
if we explain prati-mai as arising from the regular phonetic com- 
bination of prati-md and the case-ending e, we may also explain 
pra-mi, as arising from pra-md-\-i. 

The infinitives prakhyai, vik/iyai, parddai, pratimai, prayai 
(2), dyai, avayai, upayai, avasai (Delbrttck, Verbum, p. 221), to 
which must be added vayodhai, x.55.1 ; 67.11, cannot be regarded 
otherwise than as datives, although we should expect prakhy'e, 
etc. The case-ending e is added to the full root-stem : and if we 
take prame as a loc, and pratidhd'm as an infinitive, these two 
w T ords, along with mapatyai and mahiyai (p. 359 top), are the 
only formations which are in this respect analogous to them. 

B. Here belong : atithi-gvd'ya, 4 ; dyu-kshd'ya; patam-gd'ya; 
rathe-shthd'ya; sva-jd'ya (from svajf), AV. vi.56.2. 



448 G. R. Lawman, [Ead. d, d-stems. 



Dative Singular Feminine. 



A. For kshk, iv.3.6, Bollensen reads ukshn'e. 

B. Here belongs pra-jd'yai, 9. 

Dative Singular Neuter. 

A. No example. For the infinitives, see D.s.masc. 

B. Here belong: su-mnd'ya, 10; pards-pdya, VS. xxxviii.19. 

Ablative Singular Masculine. 
A. No example. B. Here belongs rpya-dd't, x.39.8. 

Ablative Singular Feminine. 

A. Here belong: gmds, x.22.6 ; jmds, i.157.1 : viii.1.18. So 
Gr., s.v. gmdl ; he calls the stem gdm an " unberechtigte Fiction." 
To be consistent, he ought to refer kshmds, i.100.15, to kshmd', 
not kshdm. Gr. reads apadhds, ii. 12.3 ; see I.s.f. 

B. No example. 

Ablative Singular Neuter. 

A. For this and the remaining cases of the neuter there is no 
example. 

B. Here belong : dur-gd't, 6 ; sadhd-sthdt, 4. With equal 
reason belongs here antdrikshdt; but it has been included above, 
p. 337. For the resolved forms in -oat, see p. 338 top. 

Genitive Singular Masculine. 

A. Here belongs certainly krshti-prds (dadhikrd'vnas), iv.38.9. 
Gr., Wb., takes papu-shds, v.41.1, as G.s.m. ; but in the Ueb., as 
N.s.m. Some recognize in jd'spati, i.185.8, a genitive jd's (from 
jd'). 

B. Here belong : a-jdsya, 2 ; atithi-gvdsya, 3 ; sva-jdsyarf, AV. 
x. 4.10,15 ; sdma-gdsya, Ait. Br. ii.22. 

Genitive Singular Feminine. 

A. Here belong : gmds, 4 ; jmds, 4. The supposed genitive 
g(a)nd's of gnd's-pdtis, ii.38.10, has the same anomaly as jd's, 
above. 

B. Here belong : jyd'yds, 2 ; prajd'yds, 3. 

Genitive Singular Neuter. 

A. No example. B. Here belong : su-mndsya, 3 ; asthvjdsya 
(kild'sasya), tan-A-jdsya, AV. i.23.4. 

Locative Singular Masculine. 
A. There is no example. 



L.s.m.] Noun- Inflection in the Veda. 449 

B. Here belong 12 forms (from 7 stems) : ddpa-gve; ndva-gve; 
ratha-samg'e (already counted, p. 338) ; sam-sth'e, 5 ; sam-ge, 2 ; 
su-khe; suar-ge. The patronymic dtithigvb (2) should have been 
counted, p. 338 ; but it was omitted. 

If we refer ufdndm to a stem uf&nd, we shall at least have to assume another 
stem, ufdnii, for the locative in i.51.11, ufdne kdvye. This is a poor make-shift. 
See as-stema, L.s.m. 

Locative Singular Feminine. 

A. There is no example. 

B. Here belong: dpa-yd'ydm; puro-dhd'ydm, AV. v.24.1-17. 

Locative Singular Neuter. 

A. No example. B. Here belong 40 forms (from 7 stems) : 
karanja-he; kh'e, 2 ; dur-ge, 5 ; bhayd-sthe; sadhd-sthe, 23 ; su- 
mne, 7 ; mahd'-vailasthe (wrongly counted, p. 338). 

In the ill-preserved hymn v.41, verse 1, we read, h6 ni, vdm mitrdvarundv 
rtdydn \ div6 vd mah&h, pd'rthivasya vd, de \ rldsya vd sadasi. Gr., Veb. i.516, 
takes di as L.s.n. of da=dam, 'room, region.' 'Who now, M. and V., is 
devoted to you either in the room of the mighty heaven, or of the earth, or in the 
place of the sacrifice V Of. p. 442. 

Vocative Singular Masculine. 

A. Here belong 21 forms (from 6 stems) : rtvrpds; bh4'ri-dds 
(iv.32.20) ; pukrapHta-pds; puci-pds, 2 ; mta-pds; soma-pds, 15 ; 
rta-pds, TS. iii.2.8 1 . 

For vayo-dhas, dravino-das, see as-stems, V.s.m. 

B. Here belong: eka-ja; dyu-ksha, 2 ; from the AV.,patam-ga, 
vi.50.2 : xiii.2.30; soma-pa, i.8.3 : ii.12.3; from the T.ar., anturi- 
ksha-pra, iv.7.5. 

Vocative Singular Feminine. 

A. There is no example. 

B. Here belong: gange; prad-d7ie, 2; tiro-dhe, AV. viii. 10.28. 

Vocative Singular Neuter. 
A. B. There is no example. 

Nominative, Accusative, and Vocative Dual Masculine. 

The ending -d occurs 27 times ; the ending -an, only 7 times. 
Since both of these endings occur also with the most of the con- 
sonant stems, it will be best to give the circumstances of occur- 
rence together at the end of this article. 

I. A. Here belong 27 forms (from 19 stems) : go-pd, 4 ; 
gharmie-shthd' ; chardishrpd' ; jagat-pd' ; tanil-pd'; tapushpd'} 
dravino-dd' (utd, p. -aij, utd), v. 43. 9 ; paras-pd', 2 ; purd-jd' : 
puru-trd, viii.8.22 ; vdja-dd; fatd-rdf; puci-pd; su-gopd' ; soma- 
pd', 2 ; sti-pd"; further, at end of catalectic pddas of 7 syllables, 
where Gr. reads -ad, kakshiaprd' , i.10.3 ; go-pd', viii. 25. 1 ■ tanil- 
pd', vii.66.3 ; from AV., go-pd', v.9.8. 



450 C. R. Lanman, [Rad. d, a-stems. 

B. Gr. refers udanya-jd" to a B-stem (■$&') ; so eta-gvd (2). 
Of the first stem no other form occurs, and of the second, none 
that cannot be referred to a stem eta-gvd. I have therefore 
counted all three forms with the A-forms. 

II. A. Here belong 7 forms (from 6 stems) : go-pau; go-pati,; 
suta-pau; from AV., adoma-dhati, viii.2.18. 

B. Gr. refers the following forms to B-stems ; but they may all 
be referred with equal i-ight to A-stems, and I have counted them 
with the A-forms. They are : d-krau (stem d-krd; cf. p. 442) ; 
adhva-gati; madhu-pau; rayi-dad. 

Nominative, Accusative, and Vocative Dual Feminine. 

A. There is no example. 

B. Here belong : rdH-pe; pUrva-je; su-dsa-sthk 

Nominative, Accusative, and Vocative Dual Neuter. 
A. No example. B. Sole example : sadhd-sthe, x.17.6. 

Inst., Dat., and Abl. Dual Masc, Fem., and Neuter. 

A. B. The Rik has no example. The TS. has several instru- 
mentals at iii.2.10 1 : vdk-pd'bhydm, kratu-, eakshush-, protra- 
pd'bhydm. 

Genitive and Locative Dual Masc, Fem., and Neuter. 

A. The Rik has no example. 

B. Locatives f. : sva-dhdyos, RV. ; jarighdyos, AV. xix.60.2. 

Nominative and Vocative Plural Masculine. 

A. The ending -as is added to the stem : as, go-pd'-as, gopd's. 
The form occurs in the Rik 84 times (from 50 stems) ; and in the 
Atharvan, 12 times (from 10 stems). Enumeration : ' 

I. (rfshayo) gd's, ' singing,' AV. xii.1.39 ; — parama-jyd's, Rik 
viii.1.30. 

n. d'-tds, iii.43.6 (see p. 441) ; d-prajds; su-prajd's. 

III. rbhu-kshds, 2 ; — eta-gvds (Gr., eta-gvd,) ; — devajd's; 
purd-, 4 ; prathama-, 2 ; sva-; — rta-jnd's, 4 ; rta-jnds, 2 ; pada- 
jnd's, 3 ; — apva-dd's, 2 ; go-; dravino-; dhana-; vastra-; vdso-; 
hiranya-; — jani-dhd's; dhiyaih-, 3 ; payo-; ratna-; reto-, 2 ; 
vayo-; — kula-pd's; go-pd's, 9 ; devd-gqpds, 3 ; su-gopd's, 2 ; 
tamd-pd's; papu-; vrata-; — agre-pd's; anjas-; payas-; soma-; 
havish-; — carma-mnd's; — jma-yd's; — dhana-sd's, 2 ; 
sadd-, 2 ; sahasra-; — prthivi-shthd' s ; — rathe-; harmie-; — 
ghrta-snd's. To these add : pravdie-jd's, raghu-jd's, vdy-d-gopds, 
— forms which Gr. refers to B-stems. 

Gr. proposes : indra-gopaas, devd-gopaas, viii.46.32 ; go-pdas, 
viii.31.13 ; su-gopdas, v.38.5c and d. These resolutions are all at 
the end of catalectic pddas of 1 syllables and needless ; read -ds : 
so in x.65.145, rta-jnd's (11 syllables — Gr., -aas, 12). 



N.V.p.m.] Noun-Inflection in the Veda. 451 

From the Atharvan : agni-jd's, apsu-, oshadhi-, x.4.23; khala-, 
viii.6.15 ; prathama-, x.7.14; pakadhu-ma-, viii.6.15 ; — vara- 
dd's, xix.71.1 ; — go-pd's, vii.81.6 : x.8.9 ; tanu-, vi.41.3 ; — 
soma-pd's, v.25.9 ; s6ma-pds, xviii.3.64. 

Resolution is necessary in RV. x.23.6, go-pdas. 

B. Here belong : ddpa-gvds ; ndva-gvds, 7 ; patam-gd's, 2 ; 
su-gd's ; and from the AV., ndva-gvds, xviii.3.20 ; tanH-jd's, 
vi.41.3 ; tapo-jd's, vi.61.1 ; vala-gd'sf, xix.9.9. 

There are four forms in -dsas. I am uncertain whether A-stems 
may form their plural thus ; whether, for instance, priya-sd'sas 
may be referred to a stem -sd'. The forms are : ddpa-gvdsas; 
dyu-kshd'sas; ndva-gvdsas ; priya-sd' sas. 

Nominative and Vocative Plukal Feminine. 

The forms of series A and B are here coincident. The Rik has 
28 forms (from 18 stems). 

Enumeration : gnd's, 6 ; vrd's, 4 ; d-gopds; ava-sthd's; dhi- 
gopds; giri-jd's; jarani-prd's; devd-gopds; deva-yd's; ndva-gvds; 
pra-jd's, 3 ; prathama-jds; pra-yd's; vakshane-shthd's; vdja-dd's; 
su-gopd's; soma-pd's; svayam-jd's. 

From the Atharvan: mamishya-jd's, xi.4.16 ; rta-jnd's, iv.2.6 ; 
pada-jnd's, vii.75.2 ; pushma-dd's, xix.40.2. 

In four instances gnd's is dissyllabic ; if gan-d's is the correct 
resolution (cf. Zend ghendo), the d may be regarded as suffixal 
(cf. p. 355). 

Nominative and Accusative Plural Neuter. 

A. I have no evidence that the following forms are to be 
referred to A-stems. 

B. 1. Here belong 15 forms (from 5 stems) : dnsa-trd (wrongly 
included on p. 346) ; dur-gd', 2 ; sadhd-sthd, 3 ; su-gd', 5 ; su-mnd', 
4 ; from AV., su-gd', vii.97.4 ; prathama-jd' r-, p. -jd', viii.9.16,21. 

2. Here belong 16 forms (from 6 stems) : ararin-ddni; khd'ni, 
4 ; dur-gd'ni, 5 ; sadhd-sthdni, 2 ; su-gd 'ni, 2 ; su-mnd' ni, 2 ; from 
AV., dur-gd'ni, vii.63.1 : ix.5.9 : xix.50.2. 

Accusative Plural Masculine. 
A. There is no certain example. Paradigm : Qankha-dhmti! s. 

1. Ludwig takes papu-sh&'s with vd'jdn, ' viehgewinnende 
krafte,' v.41.1 ; it is rather a N.s.m. of series B, p. 444. Among 
the forms there given as nominatives, there may be some accusa- 
tives p.m. in -&'s. 

2. There is no certain example of an A.p.m. of series A ending 
in -<2's. Possibly there is one in x.79.7, vane-jd's (dpvdn) ; see 
BR. vii.1800. Otherwise, we must take it as N.s.m. — sc. agnis. 
Gr. takes eva-yd's (marutas) as Ap.m., v.41.16 ; but it is to be 
pronounced as 4 syllables — eva-yd'as or -yd'vnas (cf. ii.34.11). In 
i.l43.8,/<fs is A.p.f. in form, but of common gender in meaning. 

3. Gr. refers tapo-jd'n (dpi, p. -jd'ri), x.154.5, and prta-pd 9 n 
(a-, p. -pd'n), x.27.6, to stems -jd' and -pd'. This is scarcely 
admissible ; we must assume transition-stems -jd' and -pd'. 



452 C. B. Lanman, [Rad. d, a-stems. 

B. Here belong: go-pd'n (vi.51.3) ; patam-gd'n; su-gd'n, 3; 
tapo-jd'n and prta-pd'n (above) ; and from the AV., dur-gd'n, 
xiii.2.5 ; tapo-jd'n, xviii.2.15, 18 (=r Rik x.154.5) ; soma-pd'n, 
ii.35.3 ; from the TS., pUrva-jd'n, saha-jd'n, iii.5.3. 

Accusative Plural Feminine. 

A. There is no example. 

B. Here belong 27 forms (from 15 stems) : kshd's, 2 ; gnd's, 2 ; 
jd's, 3; dn-agnitrds; anu-shthd's; apoa-dd's; devayd's; pari- 
shthd's {x.%1 . 10); prarjd's,1; maryd'ddsf; vi-shthd's; sahd-gopds; 
su-gd's; \sva-dhd's, 3 ; Isva-dhd's; from the AV., pari-jd's, xix. 
56.6 ; manushya-jd's, xii.4.43. The SV., i.323, reads rd's, a bad 
variant of Rik viii.85.13. 

For prati-shthd'\s\ (Gr.), see Ls.f. ; for gnd's as dissyllable, 
v.43.13, cf. N.p.f. ; for kshd's as dissyllable, iv.28.5, see stems in 
m, A.p.f. The resolution at the end of x.97.10a, pari-shthdas, is 
needless. 

Instrumental Plural Masculine. 

A. Here belong : agre-pd'bhis and rttc-pd'bhis (or f.), iv.34.7 ; 
go-pd'bhis. 

6r. — Wb. 175, s.v. d'td — assumes the ending -ais for series A. 
This view I cannot accept. The only forms which occur to sup- 
port it are d'tais, ix.5.5 : VS. xxix.5, and dhana-sais, x.67.7. 
These I refer to the stems d'-td and dhana-sd'. The form vayo- 
dhais, x.55.1 ; 67.11, rests on a mistake ; the text has vayo-dhai. 

B. 1. Here belong : tuvi-gr'ebhis; mithdsavadya-pebhis; ratna- 
dh&bhis, 2 ; su-gibhis, 3 ; sdma-gifohis, AV. ii.12.4. 

2. Here belong : ddpa-gvais, 2 ; ndva-gvais, 2 (and AV. xiv. 
1.56)"; patam-gats; su-khais; d'-tais and dhana-sais (above). 

Instrumental Plural Feminine. 

A. B. The forms are coincident. The Rik has 29 forms (from 
6 stems) : ratna-dhd'bhis ; gnd'bhis, 5 ; pra-jd'bhis, 6 ; $rad- 
dhd'bhis; lsva-dhd'bhis, 15 ; 2sva-dhd'bhis. 

Instrumental Plural Neuter. 

A. No example. B. 1. Here belong: madhu-p'ebhis (i.34.10: 
iv.45.3) ; su-g'ebhis, 5 ; su-mn'ebhis, 4. 
2. Here belongs su-mna'is, 7. 

Dative Plural Masculine. 
A. No example. B. Sole example : ptirva-j&bhyas (Gr., -bhias). 

Dative Plural Feminine. 

A. B. Sole example: pra-jd'bhyas, 3; prajd'bhyas (Gr. -bhias), 
at end of iv.53.4c. 

Dative Plural Neuter. 
A. B. There is no example. Cf. pp. 350, 397, 417. 



Ab.p.m.] Noun- Inflection in the Veda. 453 

Ablative Plural Masculine. 
A. Sole example : bkiXri-dd'bhyas. B. No example. 

Ablative Plural Feminine. 
A. B. There is no example. 

Ablative Plural Neuter. 
A. B. There is no example. 

Genitive Plural Masculine. 

A. No example. B. Here belong : sdkam-jd'ndm; dvi-jd'ndm, 
AV. xix.71.1. 

Genitive Plural Feminine. 

A. No example. B. Here belong : vi-bhd'ndm; pra-jfid'ndm, 
AV. xiii.2.2. 

Genitive Plural Neuter. 

A. B. There is no example. 

Locative Plural Masculine. 

A. Here belongs d'-tdsu (or f.), 2. 

B. Here belong : drwoino-deshu; su-kheshu. 

Locative Plural Feminine. 

A. B. Here belong: kshd'su, 2; gnd'su; jd'su; pra-jd'm, 2; 
sa-bhd'su. 

Locative Plural Neuter. 

A. No example. B. Here belong: dur-geshu; sadhd-stheshu ; 
au-g'eshu; su-mn'eshu, 3. 



STEMS IN RADICAL I, I, U, U, B. 

It is doubtful whether there are any stems ending in a really 
radical I or u ; cf. pradhi, p. 367, and abhidyu, p. 401. Most 
roots ending in these vowels, as well as those which, according 
to the Hindu authorities, end in r, take an added t when used 
as the final member of a noun-stem : thus, a$va-ji-t, soma-sut, 
loka-hr'-t. The stems of this sort are given by Gr., Wb. 1727-8. 
I have treated them under the head " stems in t." Of. also p. 
419 (star, tdr). 

The stems in radical long i, as deva-vi', pada-ni', manyu-mV , 
are treated with the other 2-stems (see pp. 368, 369, 375-400), 
and the forms are enumerated in each case under " C." In like 
manner, the forms of the stems in radical long d, as uda-pfi,', 
nabho-j'Q/, vtra-sA', surd-<$', are given under " C," pages 405-419. 

VOL. X. 61 



454 C. R. Lanman, [Kadical 

The following sections treat of suffixless or radical stems 
ending in consonants. The arrangement is based on the alpha- 
betical order of the final consonants. 



STEMS IN RADICAL GH. 

There is, to my knowledge, only one stem in gh in the Veda : 
sardgh, f., ' bee.' This occurs in the N.p., sardgho madhu-kr'tah, 
Qat. Br. iii.4.3". From sardgh are formed : N.s., sardt, TS. 
v.3.12 3 : <>t. Br. xiii.3.1 4 ; D.p., sardd-bhias, RV. i.112.21.' The 
d (t) is here the regular representative of Indo-European gh* ; 
compare Huhschmann, in Kuhn's Zeitsch. xxiii.386. It is then 
unnecessary to set up a stem sardt or sardh; and the stem sardgh 
is supported hy iydm . . sardghd . . . sdraghdrn mddhu, TBr. 
iii.10.10 1 . See A. Weber, Ind. 'Stud, xiii.110. 



STEMS IN RADICAL AC OR AffC. 

The root ac appears in a number of compounds which show 
such peculiarities of declension that it is necessary to treat of 
them by themselves. After them will be treated all other radical 
stems in c. 

The stems exhibit the distinction of strong and weak forms, and 
the strong forms are characterized by nasalization : thus, prd'c 
{prd-ac) makes prd'ncas in the N.p.m., but prd'eas in the A.p.m. 

These stems are declined only in the masculine and neuter. 
The feminine is formed by adding i to the weak stem of the mas- 
culine, and is declined according to series B, p. 366. 

The following lists include, I believe, all the Vedic stems whose 
final member is the root ac. First is given the element with 
which ac is compounded ; then, the stem of the masculine and 
neuter forms ; and last, the stem of the feminine forms. A stem 
is enclosed in parentheses if it has no forms occurring in the Rik 
or Atharvan. 

1. Most often the element with which ac is compounded ends 
in a or 4. The two vowels are then fused. 

Enumeration : dpa, dpdc, dpdci; ava, dvdc, dvdct; pard, pdrdc, pardd; pra, 
prd'c, prd'ci; su and pra, suprd'c, (suprdci, VS. iv.19); arvd, arvd'c, arvd'ci; 
ghrtd, ghrtd'ci ; devd, devd'ci ; asmatrd', asmatrd'c ; satrd', satrd'c, sat/rd'ci ; 
ddhara (accent), adhard'c, (adhard'ci) ; so vicva (accent), vicvd'ct. From ndra, 
instead of nard'ci, we have with irregular accent nardct', as shown by AT. v. 31.4, 
nardcid'm, ' a certain plant.' If nard'ci also occurred, we might see in nardct' 
differentiation hy means of accent, as on p. 368 top. The stem vishvd'c, of the 
isolated G.s.m. vishvd'cas, Gr. refers to vishva=vishu ; cf. vishva-dri-ak. For 
viskun&h, vr'tkak, and ninik, see A.s.n. 

If the root ac is preceded by an element ending (2) in i, or (3) 
in u, this i or u either remains intact or is changed to y or v in 
the strong cases in the samhitd, while in the weak cases it unites 
with the a of ac to i, or to H respectively. 



ac-stems.] Noun- Inflection in the Veda. 455 

2. Enumeration: (a) prati, prati-ac, praUci' (of. siprattci, VS. iv.I9); dddhi, 
dadhi-ac ; tiri, tiri-Ac (the weakest eases are supplied by a different compound, 
*tirds-ac, tirdfc); fvit-i, cvity-dc, fvitici' ; sam-i, samy-ac, samtci' ; (b) ud-i, (the 
strong eases are formed regularly from '&dac,) udici; ni, ni-ac, nt'ci; devadri, 
devadri-ac, devadri'ci ; sadhri, sadhri-ac, sadhri'ci; kadri, kadri'ci; cf, asmadri-ac, 
madri-ac, vishvadri-ac ; and tvadri-ac, madriadri-ac, yuvadri-ac (see A.s.n.) ; 
akudhri-ac ; nini-ac (see A.s.n.). 

3. Enumeration: (a) arm, anu-dc, an&ct' ; rjii, rju-ac; si,, su-ac; (b) vishu, 
vishu-ac, vish&ct. The two feminines pur-Uci' and wCtci' are formed as if from 
pwu-dc and wu-ac. Such m. stems do not occur. The place of the latter is 
filled by a compound of the root vyac, uru-vyac, whose feminine, if formed after 
analogy of samy-ac samtci', would be wru-vici'; and of this, as BR. suggest, ur&ci' 
may be a shortened form. But since uru-vydc has a regular feminine (see rad. 
c-stems, D.s.f.), it is perhaps better to refer wtu&' to uru-dc. 

4. Sometimes ac is compounded with elements ending in a 
consonant. These are : pv'it, sdm, ud, tirds, pas, r'dhf, pr'thf. 

In the first two instances, an i is inserted between the parts of the compound, 
and this i (y), or its resultant with a, goes through the whole declension : thus, 
fvity-ancas, fvitictr, cvitici'; samy-ancam, samici'. The stem ud-ac is the only one 
whose strong forms show aflc preceded by a consonant (M-arlram). Even this 
stem forms its weakest cases as though the strong were from udy-anc, and makes 
udtcas. Traces of the older formation remain in uccd', an I.s.n. with adverbial 
accent for ud-(a)c-d; cf. pafcd', for pas-{a)c-d (p. 331 above). In like manner 
tiraccd' stands, with adverbial accent, for Mras-(a)c-d ; so tirafci for tiracci. The 
feminine stem is tiracci ; but the nomen proprium, Tirafci', is differentiated by its 
accent (p. 368 top). Gr. places here, finally, r'dk-ak and pr'th-ak. 

In the Rik, if the stem is accented on the root syllable, the 
accent remains there in the strong cases (prd'ncam, pratyancani) ; 
and also in the weak (prd'ci, arvd'cd, satrd'cd), except when the 
vowel of ac unites with a preceding i or u to % or u. This 
happens with the oxytone stems of 2(a) and 3(a), and the ending is 
then accented (prattcds; prattci', N.A.d.n. — weak; dadhice, -ds; 
foittc'e; anilcds). 

In nicd' and prdcd' we have only apparent exceptions ; they 
stand for ni'cd and prd'cd with adverbial shift of accent. 

If the weak case-forms of the masculine are oxytone, the 
feminine stem is also oxytone, and this in turn throws the accent 
forward to the case-ending as stated on p. 375 top : thus, anticds 
(A.p.m.), anUci' ; prattcds (Ab.s.m.), pratipi' ; so samicy6s, 
nardcyd'm. 

These rules do not hold for the later texts. Thus the Atharvan has catr&n | jahi 
praM'co an&'cah, iii.1.4, against the Bik-reading jahi pratico armcdh, iii.30.6. See 
also A.p.m. In like manner we have sami'cas , . pacik'n, TS. v.2.9 4 ; cf. dadhi'cd 
and BR. iii.506. 

In the feminine we have : anH,'ci, AV. x.10.10 ; samt'ci, TS. iv.1.3 2 ; sami'cyd, 
TBr. i.4.8 6 ; prati'ci, AT. often; prati'cim, AV. xii.1.34: TBr. i.4.4 5 ; prati'cis, 
TS. v.2.10 2 . Like the Atharvan, the TBr.. in quoting from the Rik, modernizes 
the accent: thus TBr. ii.8.T 5 writes prati'ci for praUci' of its original, Rik i.95.5. 

The VS., TS., and TBr., however, in quoting from RV. i.96.5 and iii.29.13, do 
not alter the accent of samici', -ci'h. 

The verse RV. x.18.14 is shown to be a later addition (1) by its contents, (2) 
by its metre, and (3) by the form of transition from A to B (exceedingly rare with 
fi-stems— see pp. 401 top, 403 med.), iskv-ds. This evidence is beautifully con- 
firmed by so minute a detail as the accent of prati'cim. The examples show that 
this is a modern form; a genuine Rik- verse would have pratici'm. 



456 C. B. Lanman, [Bad. ac-stems. 

Nominative Singular Masculine. 

The case-ending -a does not appear after the double consonant 
of the strong stem -anc(-s), -ank(-s). Progressive assimilation 
reduces the form -ank to -aAn; this is preserved in the text 
before a vowel, if the thematic a is not long (as is the case with 
the stems of class 1, p. 454). 

1. The following non-assimilated form appears: pratydnk, before 
s6mo, TS. i.8.21 ; before sddo, vi.3.1" ; before shadahd, vii.4.2 6 . 
See Whitney's note to TPr. v.32. 

2. The sarhhitd has the following assimilated forms : dadhidnn, 
before a-, RV. ix.108.4; before r-, vi.16.14; niann, before u-, 
iv.13.5 : x.27.13 ; 142.5 ; pratidnfi, before a-, i.144.7 : x.79.5 ; 
before u-, i.50.56." V.2S.1. For other examples, see Ath. Pr. iii.27 
and TPr. ix.18. Cf. also ytififi for y&nk, stem y&j; ktdr'nA and 
aadr'tin (rad. p-stems, N.s.m.). 

3. Including the words just given under 2, the pada has 47 
forms in -n (from 8 stems) : dpdA; arvd'n, 23 ; dadhidn, 3 ; -ydn, 
2 ; nian, 3 ; pratydn, 12 ; prd'fi; viahvan; suprd'n. From Ath- 
arvan verses, I have noted 39 forms (from 10 stems) : adhard'n, 
v.22.2 : xii.2.1 : arvd'fi, 12 times, as iii.2.3 ; tidan, v.4.8 : ix.7.21 ; 
tirydA, thrice, as x.2.28 ; nydn, v.22.2 ; pdrdn, thrice, as vi.29.3 ; 
pratydn, 9 times ; prd'A, 5 times, as iii.4.1 ; vishvan, xi.8.33 ; 
aadhrydii, vi.89.2. 

Nominative and Accusative Sinotjlab Neotek. 

The N.A.s.n., as a "middle case" (mittlerer casus), is not nasal- 
ized, nor does it show the fusion of ia to I. All the forms here 
given are used as adverbs, unless the contrary is stated. Here 
belong 159 Rik-forms (from 22 stems) : 

1. (a) dpdk, 5 ; drvdk, as adj., i.9.5: vii.27.3 : x.29.3 ; as adv., 
39 times, and viii.50.1 ; prd'k, 6 ; from the AV., adhard'k, 
xx.128.2; 134.1-6; dpdk, xx.128.4; 134.1-6; arvd'k, 7 times, 
as iv.25.6 ; pdrdk, x.1.16 ; prd'k, 7 times, as xx.128.1. 

(5) The vowel of the final syllable -dk is shortened to -dk : 
viahund'kj vr'thdk, 2 ; *ninidk (ninio-ak — but see rad. J-stems, 
introductory paragraph). 

2. (a) pratydk; samydk, 5 ; niak, 7; aadhtiok, as adj., i.108.3 : 
iii.31.6 ; as adv., 6 times ; asmadriak, 8 (N.p.m. -driailcaa) ; 
madriak, 2 (A.s.m. -driaficam) ; vishvadriak, as adj., vii.25.1 ; 
akudhriak; from the AV., tirydk, x.2.24 ; pratydk, 16 times ; 
samydk, xviii.4.11. For samydt, see rad. J-stems, L.p.f. 

(b) The final syllable 4ak is shortened to 4k : tvadrik, as adj., 
x.43.2 ; as adv., v.3.12 (in both instances the metre allows the 
pronunciation tvadriak) ; madrik, 3 times (madriak actually 
occurs) ; madriadrik; yuvadrik; ninik, for *ninMk, — see 1(b). 
An analogous shortening is perhaps seen in jidk (16 times), for 
di6-ak. 

3. vUhvak, as adj., vii.34.13 : x.36.9; as adv., 12 times; AV., 
thrice. 



N.A.s.n.] Noun-Inflection in the Veda. 457 

4. tidak, 4 (and AV., 7) ; r'dhak, 13 ; rdhdk, once ; pr'thak, 13 ; 
praprthdk, AV., thrice, as xi.1.27. C£, however, Ind. Stud. 
iv.412,248. 

In vii.25.1, md' te rndno vishvadriag vi cdrit, vish- may be taken 
as adj. with mdno, or as adv. with vi cdrit. The example shows 
clearly how the two constructions blend logically. 

Accusative Singular Masculine. 

Here belong 33 forms (from 9 stems); arvd' fleam, 11; udan- 
eam; niancam; pratidficam, 5 ; pratydncam, 2 ; prd'ncarn, 7 ; 
madriafieam; vishuancam; samydficam; sudncam, 3 ; and from 
the Atharvan, adhard'neam, 32 times, as v.22.3 ; anvdflcam, 
vi. 134.3; dpdncam, iii.3. 7; arvd'ncam, v.3.11: xi.3.32-49 ; pdrdfl- 
cam, 21 times, as vi.65.1 ; pratydncam, 23 times, as vii.40.2 ; 
vishvancam, ii.33.7; samydficam, xiii.3.20; from TS., tirydflcam, 
ii.5.11*. 

For uru-vydflcam, see rad. c-stems, A.s.m. 

Instrumental Singular Masculine and Neuter. 

Masculine: arvd'cd (rdthena), vii.78.1. 

Neuters: devadrt'cd(mdnasd), i.93.8; 163.12; satrd' cd (rndnasd), 
thrice. The weakest forms of prd'c should be paroxytone ; as, 
prd'ed. 

If, with BB. and Gr., we take prded' as adjective with m&nasd in iii.31.5, 
prded'hinvan m&nasd saptd viprdh, the false accent adds one more item to the 
long indictment against the hymn (Ueb. i.528). It is perhaps better to take it as 
an adverb (with adverbial accent) with ahinvan. 

Five forms, which as adjectives would be paroxytone, occur as 
adverbs with the accent shifted to the ultima : tiraped', 3 times ; 
nicd', 7 (and AV. iv.3.6) ; need', 12; paped', 8; prded', ii.26.4 : 
vii.83.1. We may presuppose an adv. *pardcd', differentiated by 
the accent from the adj. I.s.n. pdrded. 

To the apprehension of the speakers, the stems of these adverbial instrumentals 
were tirafca, nfca, uccd, paped, prded, parded. These served as the bases of new 
formations : tirafed-td, 2 ; nicd't, only once ; nieafe, not until Atharvan ; uccais, 
only once ; pafcd't, 17 j prdcais, only once ; pardcais, 10. Cf. p. 337. 

These are undoubtedly late formations. The use of the plural instrumental as 
adverb belongs to the later period of the Vedic language, and none of the plural 
instrumentals thus used has the older ending -ebhis. The adv. fdnais, so frequent 
in Sanskrit, occurs only once in the RT., viii.80.3 (with panakais). 

Dative Singular Masculine and Neuter. 

Masculines : dadhlce, 2 ; pviticL For accent, see p. 455. 
Neuter : no example. 

Ablative Singular Masculine and Neuter. 

Masculine: pratieds, i. 173.5. 

Neuter: prd'eas, ii.15.3, 'von vorne.' In Ueb. i.570, for exe- 
getical reasons, Gr. takes it as Ap.m. This view is favored also 
by the accent, which we should expect to see shifted to the 
ultima, if the word were an adverb. 



458 0. B. Lanman, [Ead. ac-stems. 

Genitive Singular Masculine and .Neuter. 

Masculines : dadhtcds, 2 ; vishvd'cas. 
Neuter : no example. 

Locative Singular Masculine and Neuter. 

Masculine: prd'ci (adhvare), viii.13.30. 
Neuter : no example. 

As adverb, with adverbial accent, occurs tira$ci, Cat. Br. 
ii.3.2" ; cf. tirapci-rdji, BR. 

Vocative Singular Masculine and Neuter. 
There is no example. 

Nom., Ace, and Voc. Dual Masculine and Neuter. 

Masculines : arvd'ncd, 5 ; sadhriancd; samydncd, 2 ; from the 
AV., dpdncau, vii.70.4 ; arvd'ncau, v. 26.12 ; vwhvdncau, xx. 
136.2 ; samydncau, v.1.5. 

Neuter : predict' (dhd'mani), ix.66.2. Observe the accent. 

Inst., Dat., and Abl. Dual Masculine and Neuter. 
There is no example. 

Genitive and Locative Dual Masculine and Neuter. 
Masculine : vishdeos, L., vii.18.6. Neuter : no example. 

Nominative and Vocative Plural Masculine. 

Here belong 22 forms (from 13 stems) : (a) rjudncas ; praty- 
dncas; pnitydneas, 2 ; samydficas ; sudncas, 3 ; and from the 
AV., tirydncas, xv.3.6 ; pratydncas, xi.7.13 ; samydficas, iii.30.3,6 ; 
further (b), from the RV., arvd'ncas, 3 ; asmatrd'neas; asmadri- 
ancas; iidancas; pdrdncas; prd'ncas, 4 ; vishuancas; sadhrian- 
cas, 2 ; and from the AV., adhard'ficas, iii.6.7 : ix.2.12 ; dpdncas, 
v.3.2 ; pdrdncas, iv.40.1-8 : xi.9.22 ; prd'ncas, thrice, as xi.6.18 ; 
vUhvancas, i.19.2 : xix.38.2. There is no instance of a vocative. 

In vii.48.1, &' vo arvd'eah Jcratavo na ydtd'm | vibhvo rdiharii nariam. vartayantu, 
Siy. takes arvd'eas as equivalent to arvd'ncas. This is possible ; but, rather than 
admit so harsh a grammatical anomaly, I would even change the text out and 
out to arvd'flcas, especially since the hymn bears traces of corruption (read 
vdjds[6] — p. 345) and late origin. ' Hither may the hither-turned powers as it 
were of you as ye journey, O ye VibhU', roll the mighty wagon.' We may con- 
sider krdtavo n& as a " suggested comparison," the tertium comparaticmis (devds) 
not being expressed ; see Gr., Wb., s.v. nd 11.2). In v.33.8d, kr&tubhis can 
hardly refer to anything else than ' steeds.' 

Transition to the i-declension. A very interesting case of a new 
formation based on false analogy is seen in the late hymn x.46, 
verse 7, pviticdyas (sdmds). The proper nominatives plural are : 
m., pvitydncas; t, pvitict's. From the feminine stem pvitici', 
however, is inferred a masculine stem fvitici', and from this the 
N.p. is formed, pviticdyas. Cf. pp. 337, 371-2. 



N.A.p.n.] Noun- Inflection in the Veda. 459 

Nominative and Accusative Plural Neuter 

There is no example in the Rik. We find : pratyanci tomdni, 
Cat. Br. x.2.1"; samyafici bhdtdni, xiv.8.14 8 ; pratyanci dirghd- 
ranydni, Ait. Br. iii.44. 

Accusative Plural Masculine. 

Here belong 22 forms (from 11 stems): (a) awdcds; praticds, 
2 ; from the AV., and'cas and prati'cas, iii.1.4 (Rik. iii.30.6, -cds); 
prati'cas,v.%.1 : vi.32.3 ; vii.108.2; from the TS., samt'cas, v.2.9 4 ; 
further (b), from the RV, adhard'cas, 2 ; dpdcas; arvd'cas, 3 (for 
vii.48.1, see N.p.m.) ; dvdcas ; tidicas, x. 131.1 ; pdrdeas, 5 ; 
prd'cas, 2; vishtlcas, 3; satrd'cas; from the AV., ni'cas, xi.1.6; 
pdrdeas, ii.25.5 ; vishtlcas, iii.1.5 ; 2.3 : TS. v.2.9 4 . 

For the accent, see p. 455. Only once does the AV. have pra- 
ticds (viii.3.6), and that in a quotation from Rik x.87.4. 

Inst., Dat., Abl., Gen., and Loc. Plural. 

Of the remaining cases of the plural m. and n. there is no 
example in the RV. 

The forms are enumerated in full in all the sections on 
radical or. suffixless stems. The entire omission of a case 
signifies that no example of that case occurs. Thus no men- 
tion is made of the oblique dual cases of stems in radical dh, 
because no examples of such stems in such cases are found. 

Accent. Monosyllabic stems accent the ending in the weak 
cases. The A.p.m.i is treated in respect to accent as a strong 
case (e. g. ruc-as) ; but numerous exceptions occur, where the 
Ap.m.f. is oxytone (e. g. vdc-as). These will be grouped 
together at the end of the article. 

STEMS IN RADICAL O. 

The forms are alike for the masculine and feminine. All mono- 
syllabic stems used as substantives are feminine ; but tvac-i occurs 
twice as m., and arc-d' once as an adjective in the Ls.m. The 
only neuter form is dpr'k, an A.s.n. used adverbially. 

These stems show the distinction of strong and weak forms to a 
limited extent : 1. by nasalization (only in the stem uru-vydc) ; 2. 
by lengthening an interior radical vowel. This occurs in the 
compounds of -vd'c with drogha-, ranya-, satya-; and of -sd'c with 
apatya-, abhi-, drona-, dhdma-, nr-, yajfla- (in d-yajna-sac), rayi-, 
rdti- (and in smdd-rdti-shac), hari-. 

Nominative Sinsular Masculine and Feminine. 

Masculines : upapr'k; yatdsruk, 2 ; Mranyatvak; from AV., 
sti'ryatvak, ii.2.2 ; anrta-vd'Jc, iv.16.7. See also V.s.m. 

Feminines : tvdh; vd'k, 9 ; sr&k, 2 ; from AV., r'k, xi.7.5 : xiv. 
2.71 ; nimrtik, iv.3.6 ; vd'k, 14 times, as ii.12.8 ; cuk, xii.5.34 (and 
VS. xxxviii.18); sr&k, ix.6.17; from TS., anhomulc, ii.4.2 1 . 



460 C. R. Lanman, [Bad. c-stems. 

ACCUSATIVE SINGULAR MASCULINE, FEMININE, ANT NETHER 

Masculine. Here belong 24 forms (from 17 stems). Enumera- 
tion : (a — with strengthened stem) um-vydftc-am; — satya-vd'cam, 
2 ; apatya-sd' cam, 3 ; abhi-shd'cam ; drona-sd'cam ; dhdma- 
sd'cam; rdti-shd'cam; (b) anhomUcam; ddroghavdcam, 2 (and 
AV. vi.1.2) ; kHyavdcam; tawdriicam; dhdnaream; purordcam; 
madhupr'cam; mrdhrdvdcam, 2 ; surdcam, 3 (and AV. iv.1.2) ; 
suvd'cam. 

Feminines (91 forms, from 7 stems) : dsioam, 2 ; r'cam; tvdcam, 
12 ; vd'cam, 71 ; vd'cam-vdcam; vim/dcam; steam; sDJriatvacam; 
from AV., r'cam, vii.54.1,2 ; pdcam, iv.38.4 ; srticam, xi.1.24 ; 
tvdcam, 9 times ; vd'cam, 27. 

Neuter : dpr'k, as adverb, x.89.14 — see above. 

DjTSTBUMENTAL SINGULAR MASCULINE AND FEMININE. 

Masculines : arcd' (mdsd') ; uru-vyde-d; s&'riatvacd, 2. 

Feminines (44 forms, from 9 stems) : red', 12 ; tvacd'; mrcd' ; 
rued', 6 ; rued -rued; vded', 12 ; pucd'-pued; sicd'; surued; srucd', 
6 ; from AV., red', 2 ; tvacd', 3 ; vded', 21 ; pued', 2 ; srucd', 2. 

Dative Singular Masculine and Feminine. 

Masculines: ■Adyatasruce; yatdsruce; vipvapuce. 

Feminines : tuck, 3 ; ruce; tvace, AV. xi.2.5 ; rce, VS. xiii.39 ; 
uru-vydc-e (sdrasvatyai), AV. vi.41.2. In view of the last form, 
it is better to refer urtici', the supplementary feminine stem of 
uru-vydc, to uru-dc — see p. 455. 

Ablative Singular, Feminine. 
Here belong : tvaeds; nimrucas, 2 ; from AV., tvaeds, v.14.3 : 
vi.21.1 : xii.3.53 ; srueds, xii. 4.34. 

Genitive Singular Masculine and Feminine. 

Masculines : pururucas; surdoas. 

Feminines: reds; tvaeds; vdeds, 21; vdcas (pate); vimucas 
(napdi), 2 ; from AV, reds, ix.10.19 ; vdeds, 10 times, as i.1.1 ; 
vdcas (pate), 4 times. 

Locative Singular Masculine and Feminine. 

Masculine: tvacif, ix.69.3 (dvye) ; 101.16 (gdvye). 
Feminines : dtuci; udr'ci, 2 ; tvac't, 7 ; nimruct, 2 ; vdet, 3 ; 
vivdei, 4 ; sruci; from AV, udr'ci, vi.48.1-3 ; tvaci, i.23.4. 

Vocative Singular Masculine. 

Here belongs : dkrttaruk, x.84.4. 

Nom. Ace, and Voc. Dual Masculine and Feminine. 

I. Masculines : tawdrucd; yatdsrucd, 2 ; suvd'ed. 
Feminines: (a— with strengthened stem) satya-vded (-vdc), 

x.12.1 and AV. v.1.9 ; (b) ghrtapr'cd, RV, once ; tan-drucd, once. 

II. Feminines : vd'eau; sicau, 2. 



N.V.p.m.f.] Noun-Inflection in the Veda. 461 

Nominative and Vocative Plural Masculine and Feminine. 

Masculine. Here belong 41 forms (from 18 stems). Enumera- 
tion : (a — with strengthened stem) abhi-shd'cas, 3 ; d-yajna-sdcas; 
drogha-vd! cas ; nr-shdcas, voc. ; nr-shd'cas, 2 ; ranya-vd'cas; rayi- 
shd'oas ; rdti-shd'cas, 8 ; satya-vd'cas ; smdd-rdti-shdeas ; hari- 
shd'cas; from AV., abhi-shd'cas, xviii.4.44 ; rdti-shd'cas, xviii.3.20 ; 
(b) dudhrd-vdcas; yatdsrucas, 8 ; vasurucas; v'wdcas, 2 ; sttrucas, 
3 ; suvd'cas, 3 ; sdhtdvdcas; silriatvacas, voc. ; from AV., surticas, 
xviii.3.22 ; suvd'cas, ii.5.2 ; samsicas, xi.8. 13. 

Feminines (21 forms, from 8 stems) : dpr'cas; r'cas, 5 ; divo- 
r-dcas; pr'cas; bhadravdcas, voc. ; rdcas, 2 ; vd'cas, 6 ; sr&cas, 4 ; 
from AV., r'cas (N. and A), 14; vimdcas, vi.112.3 ; nimrucas, 
xiii.3.21 ; vd'cas, twice; sicas, xiv.2.51 ; sr-dcas, v.27.5 : xix.42.2 
(text srucdsf). 

Accusative Plural Masculine and Feminine. 

Masculines : anr'cas; articas; mrdhrdvdcas, 3 ; vddhrivdcas; 
vivdcas, 2. 

Feminines : mrdhrdvdcas/ rucas, 2 ; vd'cas, 4 ; surdcas; srdc- 
as; from AV., durvd'cas, iv.17.5 ; picas, vii.100.1 ; sicas, xi.9.18 ; 
10.20; suriicas, iv.1.1. 

Irregular accent. In i.113.17, vdcds (Hd iyarti vdhnih) must be 
an ace. (cf. ii.42.1) ; so srueds, AV. xviii.4.2. 

Instrumental Plural Feminine. 
Here belongs rgbhis, ii. 35. 12: AV. x.1.12 : xii.1.38; 4.49. 

Ablative Plural Feminine. 
Here belongs rgbhyds, AV. x.5.30 : xiii.4.38. 

Genitive Plural Feminine. 

Here belong : rcd'm, x.Tl.ll ; from AV, rcd'm, x.8.10: xv.6.3 ; 
tvacd'm, xii.3.51 ; madhupr'cdm (sc. apd'm), iii.13.5. 

STEMS IN RADICAL OH. 

Here belongs bandhu-pr'ch-d, N.d.m., iii.54.16. The forms 
vipr'cham, sarnpr'cham, sampr'che, and prchi occur as infinitives. 

STEMS IN RADICAL J. 

Several stems are treated here whose final j is not radical, but 
■belongs to the derivative suflSx aj or ij (p. 327). They are : 
trshndj (from tr'shnd), dhrshdj, sandj, upij, vanij, bhurij. Of less 
clear formation are the stems bhi-shaj and dsrj (from asar-vj? — cf. 
Cretic sap, ' blood'). The adverb ninth (which BR. refer directly 
to stem nirm) is best explained as standing for ninidk, ninid'k 
(p. 456). The ak may be referred to root ac or to the suffix 
aj (ninid-\-aj; cf. trshn&'j). 

vol. x. 62 



462 0. B. Lanman, [Bad. /stems. 

The stem dsrj is supplemented by asdn thus : dsrk, dark, asnd', 
asnds, ctsnds. 

The forms are alike for the m. and f. Most monosyllabic stems 
used as substantives are f. ; but dj, ydljf, ydj, and raj are m., and 
bhrd'j occurs as a m. adj. Neuter forms are found from the stems 
dsrj, pvdtrabhd'j, suytij, svardj, and svd'vrj. 

The distinction of strong and weak cases. 

1. Nasalization. This is seen only in two words, ytinjam (yOjam 
occurs 15 times) and yufijd (y&jd occurs once) ; but the VS. has 
also y&nn, N.s.m. 

2. Of the lengthening of an interior radical & there is no certain 
example. 

The compounds of -rd'j appear with long d in the weak eases (vird'je, samr&'je) 
as well as in the strong. In like manner we have bhrdjd' as Ls.f. Here, there- 
fore, we must consider the long vowel as inherent in the root or stem. (Different 
is the root in giri-bhrii'j.) 

The same is true of -bhd'j. The Eik has the forms : pitubhd'jas; •p'&rrva-bhd'jam, 
-bhd'jas; prathamabhd'jam ; gobhd'jas; vdmaiihd'jas ; ratnaMid'jas.Q.&.f.; fvdlra- 
Ihd'jd, I.s.n. All these, save the last two, occur in strong cases, and no form has S. 

Gr. sets up the first three stems with short U; but if the long d were not 
inherent in the stem (cf. the later forms cited by BK. v. 240), we should have to 
assume strong forms (ratndbhd'jas, fvdtrabJid'jd) in weak cases (G.s.f. and I.s.n.) ! 
Cf. dvipd'de, rad. d-stems, D.s.d. 

The verbal -y&'j, on the other hand, appears generally with & (diviyajas, satya- 
y&Jam, suyajam, prkshaprayajas — all strong forms). For ydt as a voc., see V.s.m. 
For ava-yd's, see N.s.m. 

I am inclined to think that a comprehensive examination of the 
exceptions to the law respecting the reconversion of palatals to 
gutturals before the suffix a [Lindner, p. 13, 3,] would show that 
in many cases the words with c or j belong to later texts {rued, 
VS. ; rujd, MBh.) and are by origin new forms of transition rather 
than old and normal formations (cf. -rue, rdka). 

Thus we have in viii.45.]3, indra drdhd' rid drujam; and so in iii.45.2, vrtra- 
khddo valamrtyah purd'rh darmd apd'm qjah .... indro drdhd' cid drujah. 
Despite the accent, I can not but think that these a-forms, drv/jd-m etc., are 
transition-forms made after analogy of radical stems in j (paphd-rtij-am, rtj-as, 
raiha-yij-am), and in part under the influence of the metre and the similar verse- 
ending. 

A real stem a-yuja is ill-authenticated in Vedic texts. In Par. 
Grhy. i.17.3, Codex Wilson 451 reads ayugrndksharam; and in 
Acv. Grhy. iv.2.2 ; 5.3, the edition reads a-yuj-o (N.p.). In i.15.7, 
however, we find — just where we should expect it — a form of 
transition to the a-declension, ayujdni (ndmdni), used to avoid 
the unfamiliar a-yuflj-i. Accordingly in KV. viii.51.2, ayujd 
{dsamo nr'bhih) may be due to a merely formal parallelism ; cf. 
p. 468. The metre would be good with aydg utd'samo nr'bhih, 
or the like. 

The late forms bhrdjd-s and bhrdjd'ya have bhrdj-d', perhaps, 
for their point of departure. 

Transition to the ^-declension. The form tujdye (accent — cf. 
tuj-k, iv.1.3 : v.41.9), v.46.7, is a transition-dative to t<ij, equivalent 
to Vuj'e in meaning. There is no oxytone stem tuji. 



N.s.m.f.] Noun- Inflection in the Veda. 463 

Nominative Singular Masculine and Feminine. 

The final radical,/ appears as k or as t in general according as it 
represents an original g 1 or g*. See Ascoli, Corsi di glottologia, 
i. 106,1 17 ; and Hubschmann, in Kuhn's Zeitsch. xxiii.3S4-90. 
Cf. stems in rad. h, N.s.m.f. 

I. Masculines : abhi'ik; upik, 4 ; rtay&k; rtv-ik; ghrtdnirnik, 
3 ; candrdnirnik ; pardvr'k, 2 ; bhishdk, 3 ; vanik ; sarhvr'k ; 
sayuk; from A V., ardha-bhd'k, vi.86.3 ; bhishdk, v. 2 9.1 : x.4.15 ; 
catd-bhishak, xix.7.5 ; hiranyasrak, x.6.4. See also V.s.m.f. 

Feminines : nirnik, 2 ; hiranyanirnik, 2. 

(a) Here also belongs the form yunH (asi), VS. x.25 ; it stands 
for yi'mk. Cf. p. 456 and Vaj. Pr. iv.104. 

II. Masculines : ekard't; bhrd't; rd't, 2 ; vanerd't; vibhrd't, 2 ; 
vird't, 2; vipvabhrd't; samrd't, 12; $vard't,5; from AV., ekard't, 
iii.4.1 ; vird't, or f., 20 times, as ix.10.24 ; samrd't, thrice. Cf. 
ydt, V.s.m. 

Feminines: rd't, v. 46. 8; vird't, twice. 

III. The form avayd's occurs in i. 173. 12 as f. (p. ava-yd'h) ; so 
AV. ii.35.1 (p. ava-yd'f) ; further, d'vayds, m., i.162.5 (p. d'- 
vaydh!). Cf. Ind. Stud, xiii.96. These forms are referred by 
BR. and Gr. to stems avayd'j and d'vaydj. It is doubtful 
whether the long vowel is inherent in the stem. The forms of the 
grammarians, avaydjam etc., do not occur; and the d of the nom. 
may be explained as a vicarious lengthening. The genetic series 
would be : ava-yag'-s, -yaz-s, -yac-s, -yd's. See Johannes Schmidt, 

Verwantschaftsverh&ltnisse, p. 11 ; and S. Goldschmidt, Z.D.M. G. 
xxvii.709. Cf. pveta-vd's, from -vaz-s, -vaghfs. Sadha-md's and 
puro-dd's are not entirely parallel. 

Transition to a-declension : bhrdjd-s, x. 170.3 ; see also p. 462. 

Nominative and Aoousative Singular Neuter. 

Here belong : d#rk, i.164.4 : AV. iv.12.4,5 : Ait. Br. ii.9 ; svd'vrk 
(amr'tam), RV. x.12.3; suyuk, as adv., iii.58.2 ; and perhaps 
ninik (see pp. 461, 456). For dsrt, see p. 466 top. 

Accusative Singular Masculine and Feminine. 

Masculines (61 forms, from 22 stems) : (a) ydnjam; (b) upijam, 
3 ; rtv'ijam, 1 1 ; jyeshthard'jam, 2 ; trshndjam; pardvr'jam, 2 ; 
pdrvabhd'jam, 2 ; prtdnd'jam; prathamabhd'jam; prdtarydjam; 
yujam, 15; yuvdyujam; rathayujam, 2; vird'jam; paphdrujam; 
satyaydjam, 2 ; samrd'jam, 4 ; sayujam; sutydjam; suydjam; 
suytijam; svard'jam, 5 ; svanr'jam; from AV., dj-am, xix.50.5 ; 
tribhujam, viii.9.2 ; vird'jam, or f., 6 times, as viii.9.7,9; vanijam, 
iii.15.1 ; susrdjam, xx.128.15. 

Feminines (61 forms, from 10 stems) ; upaspijam; H'rjam, 36 ; 
t'Ajam; nirnijam, 13 ; parivr'jam; bhujam, 3 ; manoyujam, 2 ; 
vird'jam; sahdsranirnijam ; srdjam, 2 ; from AV, bhujam ; 
svard'jam; u'rjam, 27 ; srdjam, 2. 



464 C. R. Lanman, [Rad. j-stems. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculines: manoyujd; yujd', 30 times, and i.39.4: viii.85.15 ; 
vacoyujd; sahdsranirnijd, 2 ; suyujd, 3 ; from AV., bhishdjd, 
ii.9.5 ; yujd', iv.23.5 : v.21.11 : xiii.1.3 ; say4jd, v.14.7 : vii.108.2. 

Feminines : "drjd', 6 ; tujd', 4 ; nirnijd; bhrdjd' ; from AV., 
manoyujd, v. 7.5 ; 10.8 ; vird'jd, xv.14.5. 

Neuter : pvdtra-bhd'j-d (vdyasd), viii.4.9. 

Dative Singular Masculine and Feminine. 

Masculines : trshndje, 2 ; rakshoyuje ; varuje ; samrd'je, 2 ; 
svard'je, 3 ; from AV., mrtd-bhrdj-e, iv.4.1 ; yuje, vi.54.1,2 ; vird'je, 
samrd'je, and svard'je, xvii.22,23. Transition-form : bhrdjd'ya, 
VS. viii.40. 

Feminines : Hrj'e, 2 ; tuj'e, 2 ; nirnije, 4 ; bhuj'e, 5 ; from AV., 
tuj'e, vi.33.1 ; iJbrj'e, twice. Transition-form: tujdye, v.46.7 (p. 462). 

Ablative Singular Masculine and Feminine. 
Masculines : yujds; vird'jas. Feminine : nirnijas. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines : upijas; samrd'jas; svard'jas, 2. 

Feminines: u-rjds, 11 ; Hrjas, with voc, 10; prayujas; yujds; 
ratna-bhd'j-as (p. 462) ; from AV., ttrjds, 4 ; vird'jas, 6 ; bhraj-ds, 
vii.90.2. 

Neuter: svard'jas (gotrdsya), x.120.8. 

Locative Singular Masculine and Feminine. 
Masculine : dpvanirniji. Feminine : samsr'ji. 

Vocative Singular Masculine. 
Here belong : ghr'tanirnik ; samrdt, 3 ; perhaps dkrttaruk, 
AV. iv.31.4 (cf. RV.'x.84.4,' c-stem). 

BR., vi.8, take ydt, x.61.21, as a verb; so S4y., aydt | ycya. Both proposed 
to take it as a vocative, ' offerer.' Since the j of yqj represents original g\ the 
normal nom. s. would be yd't ; and corresponding to this, according to the gram- 
marians, the voc. s. would be ydt (cf. prdt). 

Nom., Aoc, and "Voc. Dual Mascultne and Feminine. 

I. MaBculines (44 forms, from 13 stems) : (a) yunjd; (b) rtv'ijd, 
2 ; tanutydjd; purubhujd, 10 ; pushkarasrqjd; prdtaryujd; brah- 
mayujd, 3 ; bhishdjd, 5 ; yujd; vacoyujd, 4 ; samrd'jd, 10 ; sayujd; 
sdkamyujd; suyujd, 3 ; from AV., p&shkarasrajd ; bhishdjd; 
sayujd, 2 ; manoyujd, vi.65.1. 

Feminine : sandjd, i.62.7. 

II. Masculines : apnard'jau; bhishajau; samrd'jau; from AV., 
sayujau, vii.53.2 ; suyujau, vi. 140.3 ; awAvr'jau, ix.4.12. 

Feminine : afvayujau, AV. xix.7.5. 



D.d.m.] Noun- Inflection in the Veda. 465 

Dative Dual Masculine. 
Here belongs bhishdgbhydm, AV. x.6.12. 

Genitive and Locative Dual Masculine and Feminine. 

Gen. masc. : samrd'jos. Gen. fern.: bhur'yjus, iv.2.14: viii. 
4.16 : AV. xx.127.4. Loc. fem. : bhurtjos, ix.26.4 ; 71.5. 

Nominative and Vocative Plueal Masculine and Feminine. 

Masculines (63 forms, from 28 stems) : dtrshnajas; dsvapnajas, 
2; up'tjas, 14; rtvijas, i; krtddhvajas; giribhrdjas; ghrtdnirnijas; 
tanutydjas; trshndjas; diviydjas; durytijas; dhrshdjas; nihsrjas; 
pitubhd'jas, 2 ; ptirvabhd'jas; prtand'jas; brahmayvjas, 2 ; bhi- 
shdjas, 2 ; manoytijas, 4 ; mitraytijas; yiijas, 2 ; rathaytijas, 2 ; 
vurshdnirnijas, 2 ; vdmabhd'jas, 2 ; samrdjas ; samrd'jas, 8 ; 
suytijas, 4 ; svaytijas; svard'jas, 3 ; from AV., svard'jas; ytijas, 
2 ; bhishdjas, 3 ; rtvijas, 7 ; rtvijas, vi.2.1 ; dy&jas and prayujas, 
xi.8.25 ; vi^vasr'jas, xi.7.4 ; hdritasrajas, x.8.31. 

Feminines : abhiyujas ; u$yas ; €rjas ; gobhd'jas ; ttijas ; 
nirnijas; prkshdprayajas ; praydjas, 3 ; bh-djas ; standbhujas; 
sudyujas; from AV., ti'rjas, xii.1.12. 

Nominative and Accusative Plueal Neutee. 

No example. Paradigm : apva-yunji. In place of this, a form 
of transition is used : e. g., ayujdni (p. 462). 

Accusative Plueal Masculine and Feminine. 

Masculines : ddhinirnijas ; apvaytijas; upijas, 2 ; caturyujas; 
vijas, 2 ; $aphdri(/jas (and AV. viii.3.21) ; sayujas; mytijas, 3 (and 
AV, 6). 

Feminines : abMydjas, 5 ; tujas; prayiijas; bhujas, 2 ; srdjas 
(and AV. xx.127.3). 

Instrumental Plueal Masculine. 

Here belong : anmayugbhis ; uplgbhis, 2 ; rtayugbhis, :s ; 
suy&gbhis; svayugbhis, 2 (and AV. ii.5.4). In AV. vii.4.1<£, read 
niyddbhis (BR.). 

Dative and Ablative Plueal Masculine. 

Dative : tipigbhias. Ablative : bhishdqbhias, AV. xix.2.3 ; 
56.2 ?. 

Genitive Plueal Masculine and Feminine. 

Masculines : upljdm, 3 ; rtvljdm; bhishdjdm (and AV. vi.24.2). 
Feminines : drjd'm, 2 ; mrmajdm; bhujd'm. In 5 instances 
Gr. resolves : ttrjadm, -ddm. 



466 C. R. Lanman, [Radical 

Locative Plural Feminine. 

Here belong: sraksMt; praydkshu, AV. v.27.5,6. For pray dk- 
shu the TS. (iv.1.8 1 ) and VS. (xxvii.14) read praydtsu! Cf. dsrd 
(d-), p. dsrt, TS. vii.4.9 ; samydt te (for samydk), i.2.7 1 ; and the 
etymology: ydd ddhriyata tdd ghrtdm abhavat, ii.3.10 1 , 
cited by Weber, Ind. Stud. xiii. 106,107. 

STEMS IN RADICAL T AND D. 

No stem ends in t. For raghdt-as, AV. viii.7.24, BR. suggest, 
raghdvas (i. e., $y end's). 

Two stems end in d, I'd and id (='tsh). The former occurs only 
in the I.s.f., idd', viii.39.1. The latter occurs as follows : idd', 3 ; 
idds, as G.s., 10. In vii.47.1, Gr. takes idds as A.p. ; but the 
accent would then be exceptional. I join it as G.s. with ■drmim. 
See sA-stems. 

STEMS IN RADICAL T. 

A very considerable number of stems are treated here whose 
final t is not radical, but belongs to one of the derivative suffixes 
it, ut, vat, tdt, at (p. 327). They are : tadit, div'it, yoshit, rohit, 
sard, hard; mar&t; arvdvdt, dvdt, udvdt, nivdt, pardvdt, pravdt, 
samvdt; upardtdt, devdtdt, vrkdtdt, satydtdt, sarodtdt; veh-dt, 
vali-dt, srav-dt, sapc-dt. Isolated stand ndpdt, tdntindpdt, and 
prdnapdt. 

It will be observed that some stems admit of a double analysis: 
either as radical stems from roots ending in t; or as stems formed 
by root + consonant t. We may divide : milla-kr'-t, or -kr't (kart, 
' cut') ; vr't, or vr'-t (BR.) ; ishu-kr'-t (Mahldh. to VS. xvi.46, ishdn 
vdndn kurvanti te ishukrtah), or -kr't (Gr.). In div-it Gr. sees 
the suffix it; BR., the root i -\- t. Cf. didydt and didyd; dyvt 
and dyd. 

Three stems of exceptional formation are supplemented by 
others : ndpdt, by ndptar; ydhrt, by yakdn; pdkrt, by $akdn. 

The forms are alike for the m. and f. There are but four mono- 
syllabic stems, dytit, nr't, pr't, vr't — all fem. Six neuter stems 
are found : akshipdt and dnapdvrt; pdhrt and ydkrt; samydt (in 
I.s.) ; and trivr't, in the sense of 'amulet.' For prdt, see rad. 
«A-stems. 

No stems exhibit the distinction of strong and weak cases ; but 
the accusatives pi. vr'tas and nr'tas are accented regularly as 
strong cases. 

The stems in tdt perhaps owe their origin to forms of transition 
on the part of the £d!£«-stems to the consonant or ^-declension. For 
these transitions the L.s. might serve as a point of departure, the 
forms devdtdtd, satydtdtd, sarvdtdtd beiug easily referable to 
tf-stems (as adA*erbial instrumentals), or to t'-stems (as locatives) : so 
in x. 11 1.4 — see I.s.f. A like transition is seen in the td-stvm 
avi'ratd, which makes a D.s. from a ^stem avi'rat-e; see p. 359 
med. 



Osteins.] Noun- Inflection in the Veda. 467 

Prom sacc, : stick, stop, cease,' — i. e. sa-s(a)c, a reduplicated form of sac'f — are 
derived two peculiar formations. We have : 1. the substantive sacc-at (like 
vah-af), f., 'a stopping,' and concrete, 'a hinderer;' and 2. the participle sdfc-at 
(ii.16.4, vrshabhd'ya saccate). 1. From the noun sacc-at we have the bahuvrihi 
adj. a-saccat, with regular accent (ct Garbe, in Kuhn's Zeitsch. xxiii.512). This 
appears in the mase. form : asaccdtd, N.d., vii.67.9 ; and in the fern, forms : 
asacc&tam, As., ii.32.3 ; -aid, I.S., x.69.8 ; -atd, N.d., i.160.2 ; -atas, N.p., 10 
times. 2 . From the participle s&cc-at we have the negative compound a-saccat, 
with regular accent (of. Garbe, I.e., p. 494). This appears in the masculine 
dsaccatam, i. 112.9 (formed without nasal like participles of reduplicating class, 
dadat etc.); and in the fern, forms: Asaccantt, N.s., iii.57. 6 : riii.31.4; asaccantt, 
vi.70.2 (formed with nasal, the character of sacc as reduplicated root being 
obscured by the syncopation of &). 

Nominative Singular Masculine and Feminine. 

Masculines : andvr't; cikit, 4 ; tadit; tdn&ndpdt, 2 ; ndpdt, 14 
prdnapdt; manaccit; vipaccit, 2; vish-dvr't ; suvr't, 3; from 
the Atharvan, ekavr't, xiii.4. 12 ; dbandhuhrt (iv. 1 9. 1 ) and rmilakr't 
(iv.28.6), from ka.rt, ' cut ;' vipaccit, 4 times ; napdt, 5 ; tdnt'i- 
ndpdt. For sttyavasd't, RV. x.106.10, the padakdra writes 
suyavasa-dt ! 

Feminines : didyut, 8 ; pardvdt; pravdt, 2 ; rohit; vidyilt, 8 ; 
from AV., apacit, vi.83.3 ; vehdt, iii.23.1 ; vidytit, 6 times. 

Nominative and Accusative Singular Neuter. 

Here belong: trivr't (dnnam) ; cdkrt ; aks/dpdt (2) and dn- 
apdvrt (2), as adverbs ; from AV., ekavr't, viii.9.25,26 ; trivr't, 
v.28.4,6,11 ; ydkrt. ix.7.11 : x.9.16; cdkrt,' xil4.9. 

Accusative Singular Masculine and Feminine. 

Masculines (51 forms, from 10 stems) : acitam; ghrta-ccutam, 
2 ; tdnUndpdtam; trivr'tam, 4 ; ndpdtam, 20 ; madhu-ccutam, 
10 ; vipaccitam, 6 ; vishtivr'tam; mvr'tam, 5 ; huraccitam; from 
AV., duccitam, 2 ; vipaccitam, 3 ; ndpdtam; trivr'tam, 3 ; ekavr't- 
am, xiii.4.15. 

Feminines: arvdvdtam ; asacedtam (ii.32.3); dvr'tam, 2; 
ghrta-ccutam, 3 ; dy&tam; pardvdtam, 3 ; pravdtam; yoshitam; 
vicrftam; vidyutam, 2 ; vipaccitam; vr'tam, 2 ; samydtam, 5 ; 
samcr'tam; from AV., vipaccitam, 2 ; vidytitam, 3 ; pardvdtam, 

2 ; pravdtam; yoshitam; rohitam, iv.4.7 ; dvr'tam, x.5.31bis; 
sarhvr'tam, viii.6.4 ; vehdtam, xii. 4.37,38 ; samvdtam, vi.29 3 • 
105.2. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculines : trivr'td, 4 ; divitd; suvr'td, 6. 

Feminines (36 forms, from 15 stems) : asacedtd (x.69.8); dcitd; 
udvdtd; ghrta-ce&td; trivr'td; devdtdtd, 3 ; dyutd', 2 ; nivdtd; 
praedtd; pravdtd, 12 ; vidyi'itd, 5 ; vrtd', 2 ; satydtdtd; sarvdtdtd, 

3 ; haritd; from AV., vidy&td, 4 ; nivdtd, v.3.2. 

The forms from ^-sterns may also be taken as locatives s. of 
tdti-stema ; see p. 466. See also p. 386 top, and Wb. 1490. 



468 0. R. Lanman, [Bad. Osteins. 

Neuters: samydtd, i.151.8 : vi.16.21 ; trivr'td, AV. v.28.2,8 : 
xix.27.3 and 9 (m. ?). 

Dative Singular Masculine and Feminine. 

Masculines : acite; vipapcite, 2 ; sudynte. 

Feminines: devdtdte, 3 ; vidyute, AV, i.13.1 : xi.4.2. Here Gr., 
after Bollensen, puts ishu-kr't-e va, p. -d-iva, i. 184.3 ; see N.d.m. 
Transition from the d-declension : avi'rat-e (pp. 466, 359). 

Ablative Singular Masculine and Feminine. 

Masculine: vidytitas (haskdrd't), i.23.12. 

Feminines : arvdvdtas, 4 ; itdvdtas, 3 ; didy&tas; nivdtas, 2 ; 
pardvdtas, 35 ; pravdtas, 3 ; samvdtas; from AV., iridy&tas, 2 ; 
pardvdtas, 3; pravdtas, iv.25.6. In RV. x. 142.4, ndvdtas and 
nivdtas may be accusatives pi. 

The TS., i.8.14 1 , has didy&n md pdhi (p. didyot) ; the VS., xx.2, vidyot pdhi 
(Mahldh., vidyutah mdm pdhi). Weber would explain didyot as standing for 
didyds (stem didyu), with irregular phonetic transition of s to t; see Kuhn's 
Beitrage, iii. 388-9. He would accordingly read didyot in the VS. ; see Ind. Stud. 
xiii.101, note. Roth, on the other hand, explains vidyot as an ungrammatieal 
formation from the stem vidyfat, and as due to the parallelism of the formula, 
mrtySh pdhi vidyot pdhi (BE. vi.1066). That such outward parallelism in the 
sound of a formula could call forth such irregular formations is clear from AV. 
xvi.4.6, ushaso doshasap ca. The stem doshd' passes into the as-deelension simply 
on account of the juxtaposition of ushas ; see as-stems, C-r.s.n. A similar ungram- 
matical form is probably cakshos, stem cdkshus (p. 410 top). We need not assume 
a stem cakshu; see p. 412 end. Cf. pathe-shthd' (for pathi-), an ungrammatical 
imitation of rathe-shthd'. 

Genitive Singular Masculine and Feminine. 

Masculines : trivr'tas; vipap&tas, 2 ; sacitas; sudytitas, 2. 
Feminines: pravdtas, 2; vidytitas ; pravdtas, AV. xii.1.2 ; 
pravatas, with voc., thrice. 

Locative Singular Feminine and Neuter. 
Feminines : arvdvdti, 1 ; upardtdti, 2 ; devdtdti, 2 ; pardvdti, 
17 ; vrkdtdti. 

Neuter : trivr'ti, AV. v.28.3. 

Vocative Singular Masculine. 

Here belong: taw&napdt, 5; napdt, 17; perhaps marut in 
evayd'marut, v.87.1-9. 

Nom., Ace, and Voc. Dual Masculine and Feminine. 

I. Masculines: asapedtd, vii.67.9; napdtd, 3 times; ndpdtd, 
4 ; vipapcitd, v.63.7 (and AV. vi.97.2) ; ishu-kr'td, i.184.3 (? see 

D.s.f.). 

Feminines: asapedtd, i.160.2 ; dvr'td; madhyrpedtd ; harltd. 

II. Feminines: vr'tau, 2 ; vicr'tau, AV. ii.8.1 : iii.7.4: vi.121.3. 

Genitive and Locative Dual Feminine. 
Genitive: harUos. Locative: vicr'tos, AV. vi.110.2. 



N.V.p.m.f.j Noun- Inflection in the Veda. 469 

Nominative and Vocative Plural Masculine and Feminine. 

Masculines : dvr'tas; rshtividyutas; ghrta-pctitas, 3 ; ndpdtas, 
3 ; vidy&tas; vipapcitas, 9 ; samydtas, 2 ; sHriapvitas; svdvidyut- 
as; haritas; hrddunivr'tas; from AV., dupcitas, v. 3 1.5; ekavr'tas, 
xiii.4.13,21 ; vipapcitas. In the Rik, martitas occurs 133 times as 
N. or A., most often as N. As vocatives occur in the Rik : 
indrdmarutas; rshtividyutas; napdtas,5; marutas,176; vidyutas. 

Feminines : asapcdtas, 10 ; udvdtas; ghrta-pcdtas, 3 ; taditas; 
pravdtas; madhu-pc&tas; vahdtas; vidytttas, 10 ; samydtas, 3 ; 
saritas; sravdtas, 5 ; haritas, 12 ; frota AV., apacitas, viL76.2fo'.?/ 
dpacitas, vi.83.1 ; dvdtas, 3 times; vidyutas, 4; pardvdtas; 
yoshitas, 3 ; dvr'tas, 2 ; updvrftas; haritas, 7 ; saritas, xii.2.41. 

Accusative Plural Masculine and Feminine. 

Masculines : acitas, 3 ; ady&tas; ghrta-pcdtas; mardtas, see 
N. ; vipapcitas; hurapcitas. 

Feminines : udvdtas, 2 ; dakshindvr'tas; didy&tas, 2 ; nivdtas, 
2 ; pardvdtas, 6 ; pravdtas, 7 ; madhu-pc&tas ; rohitas, 3 ; vidydtas; 
vr'tas, 3 ; samydtas, 2 ; samvdtas, 2 ; saritas/ sapcdtas (i.42.7 : 
iii.9.4: vii.97.4) ; sravdtas, 2; haritas, 11; from AV., nivdtas; 
pardvdtas, 6 ; pravdtas, 5 ; yoshitas, 2 ; trivr'tas; sathvr'tas, 6 ; 
purUvr'tas and vishUvr'tas, x.2.11 ; nr'tas, x.2.17. 

Instrumental Plural Masculine, Feminine, and Neuter. 

Masculine : martidbhis, 35 (and AV, 7). Feminine : pravdd- 
bhis. Neuter: trivr'dbhis (gender?), AV. xix.27.3,9fo*. 

Dative and Ablative Plural Masculine. 

Dative : martidbhyas, 3 ; -bhias, 10. Ablative : marudbhias, 
i.85.8. Gr. proposes the resolution -bhias in 11 instances; in only 
one_(x.77.7) is it necessary — the others being at the end of cata- 
lectic pddas. In v.54.9, read pray ddbhy as — root i. 

Genitive Plural Masculine and Feminine. 

Masculines : vipapcitdm; marutdm, 2 ; mar&tdm, 30 (and AV., 
6) ; mari&tadni, 7 (resolution necessary in all instances but i.3S.10 
and viii.83.1). 

_ Feminines : pravdtdm, 2 ; from AV, apacitdm, vi.25.1-3 : 
vii.74.1 ; vidy&tdm, xix.44.5 ; sravdtdm, vi.86.2. 

Locative Plural Masculine and Feminine. 

Masculine : marutsu, 8. 

Feminines : udvdtsu; nivdtsu; pravdtsu; prtsti, 31. 

Double ending. Little importance ought to be attached to the 
oft-mentioned curiosity prt-sd-shu. The fact that it occurs in 
hymn i.129 (verse 4) detracts considerably from its value. 

vol. x. 63 



470 0. R. Lanman, [Radical 

STEMS IN RADICAL TH. 

There are five stems in th : pdth and rnpdth, m. ; abhipndth, 
adj. m. ; kdprth; prdth, n. The final of the last stem is uncertain. 
BR. set up $rdt or prdd; Gr. connects prdth with the verbal root 
prath. BR. and Gr. regard kdprth as masculine ; but it seems 
necessary (with A. Ludwig) to consider it as neuter on account of 
the A.s. kdprt, x.101.12. In this verse, the stem shows also a 
form of transition to the a-declension, with shift of accent, 
kaprthd-m. The stem pdth supplements pdnthd (p. 441) and 
pathi. Gr. explains the d of pdth-ds at the beginning of ii.2 Ad 
as a metrical lengthening ; this is hard to believe for the first 
syllable of the pdda. The A.p. of pdth is accented irregularly as 
a weak case, path-ds. The forms follow. 

Singular: N., kdprt, x.86.16,17; A., kdprt, x.101.12; prdt, 8 
times ; \.,pathd', 28 (and AV., 6) ; pathd'n a-, i.129.9 (cf. p. 335) ; 
supdthd, 3 times ; D., path'e, VS. xviii.54 ; Ab., pathds, 5 times 
(and AV., 2) ; abhipndthas (vdjrdt), x.138.5 ; G., pathds, 4 times 
(and AV. v.30.7) ; pathos (pate); pdthds, ii.2.4 ; L., pathi, 3 
times (and AV. xiv.1.63). 

Plural: A., pathds, 28 (and AV, 5) ; pathd'm, 3 (and AV., 2); 
pathddm, vii.73.3. 

STEMS IN EADICAL D. 

Here belong about a hundred stems, from the roots ad, kshad, 
chad, chid, trd, nid, nud, pad, bhid, mad, mud, rud, lvid, 2vid, 
sad, sud, sild, syad. The most frequent are those with -vid and 
-sdd. I have not distinguished lvid from lvid in the enumera- 
tions. In dyur-ddd-am, AV. vi.52.3, we have a new formation in 
which the reduplicated root dad (from dd) is treated as a primitive 
verbal. 

The stems in which the d is suffixal are comparatively few, and 
are : drshdd, dhrshdd, bhasdd, subhasdd, vandd, pardd. For 
kdktid, kdkiid,see p. 471. 

The forms are alike for m., £, and n., except of course in the 
N.A.n. There are 5 monosyllabic fem. stems : ud, nid, bhid, 
rrvdd, vid ; one masc, pdd ; and one neut., hr'd. The other 
neuters are compounds of pdd, and several with chid, vid, and 
syad. Hd'rd also appears in composition with dies and su. The 
compounds of pad generally form the fem. stem with t. 

The distinction of strong and weak cases is seen only in pdd 
and its compounds. The a is lengthened in the strong cases. 
There is, however, a curious wavering between the strong and 
weak stem in the N.A.s.n. ; as, dvipdt or dvipa't. Once in the 
N.p.m. we have dvip&'das, AV. In the D.s. dvipd'de, the d is 
metrical (cf. abhimdtishd" 'has, i. 9 1.1 8, and satrdsd'he) ; but not in 
■pdtsu. Of sadham,d'd and somamd'd only strong forms are 
found ; but the d is hardly inherent in the stem. 

Irregularly accented are : sdd-d, I.s. ; pad-ds and nid-ds, Ap. 

Irregular phonetic treatment of the final stem consonant appears 
only in padbhis; see Lp.m. 



easterns.] Noun- Inflection in the Veda. 471 

Transitions to the vowel-declension. Several stems exhibit 
most interesting forms of transition. Pdd is the Vedic stem for 
'foot.' Undoubtedly its strong A.s. pd'd-am was the point of 
departure for the transition to the a-declension. As if the A.s. 
were pd'da-m, the N.s. pd'da-s was formed, and the N.p. pd'dds. 
These are the only transition-forms which the Rik shows, and — 
what is more important — they occur in the latest parts of the 
whole sarhhitd. 

Pd'das occurs in the Purusha-hymn (x.90.3,4) — confessedly one 
of the very latest interpolations in the Rig? eda ; it is not N.p. of 
pdd, but a transition N.s., and signifies ' one quarter' (a late and 
derived meaning of the word — Say., caturtho 'npah, lepah), thus 
answering to tripd'd, 'three quarters.' Pd'dds, N.p., occurs in 
iv.58.3, and means the ' feet' of a melting-pot. Hymn 58 is the 
last of the book in our text, and certainly did not belong to the 
original collection. Pd'dds, N.p., occurs also in verse 9 of the 
mystical hymn to the Sun-horse, i.163. In the well-known modern 
medley, i.164, verse 12, occurs pdnearpddam, A.s.m. ; this might 
be referred to -pad; but cf. yadd . . . catushpddam (nom. s. n.) 
pancapddam . . bhavati, Adbh. Br. xii. Finally nipddd's, 
N.p.m., occurs v.83.7. 

The stem nid shows a solitary form of transition to the 
(^-declension in vi.12.6, sd tvdrh no [pdhi] aravan niddydh (BR.), 
or sd tvdrh no arvan [vi mueo] niddydh (Gr.). If nid-d'', the I.s. 
of nid, were the point of departure, the accent ought to be 
nidd'yds; cf. ndv-d and ndvdyd, p.*434. 

From sadha-md' d-am, felt as sadha-md'da-m, may be derived 
the forms sadha-md'de, -md'deshu. 

The stem hr'd is supplemented by hr'daya. Instead of hr'd 
(N.A.S.), we find hr'dayam; instead of hYndi, hr'dayd. The 
forms hr'daydt, -e, -dni, and -eshu occur almost exclusively in late 



Transition from the vowel-declension. The stem kdkuda does 
not occur until the AV. and later texts — Ab., kdkuddt, AV. 
x. 10.19 ; and unless we assume two independent formations in the 
case of kaku(b)hd and kak-ubh, I am strongly tempted to believe, 
in view of the accent, that the oxytone vowel-stem is the older, 
and that here the transition has been in the opposite direction, 
i. e. from the cf-stem to the consonant-stem. The majority of 
forms in the older texts can be referred to vowel-stems. 

The d of kdkud and k&kud is of problematic nature. Is it radical or suffixal ? 
and if not suffixal, does it represent a radical bh f 

Fick, indeed, separates kak-ud from ka-kubh, and refers the former to root leak — 
the latter to kubh ( Vgl. Wb. 3 i.36 and 51). It is, however, by no means certain 
that kakud and kakubh are not identical. Weber thinks that the dental of kakid 
is possibly a phonetic alteration of the labial. For this and other instances of 
equally anomalous changes, with suggestions respecting their significance, see 
Ind. Stud, xiii.108-110. 

"We have the following forms with the dental : N.s., kdkut, vi.41.2 ; k&kut, viii. 
44.16: AV., 4times: TS. i.5.5' : vii.2.5 2 - 3 , q.v.: VS. iii.12; kdkudam. viii.58.12; 
kdkudas, Ab., i.8.7 ; Mkudi, AY. iii.4.2 : vii.76.3 ; (iti-kakudas, TS. v.6.17*. It is 
not entirely clear whether these forms justify our setting up a stem kakud. One 
does not set up dental stems for asrt and praydtsu, p. 466. 



472 C. R. Lanrnan, [Bad. d-siems. 

Nominative Singular Masculine and Feminine. 
Masculine. There are 97 forms (from 51 stems). 

{a) The stem-vowel is lengthened. Forms : apd't, 2 ; ekapdt, 1 ; cdtushpdt, 2 
tripd't (x.90.4); dvipd't; vipvdtaspdt; sahasrapdt; sadhamd't; from AT., pd't, 
xix.6.2; apd't; cdtushpdt, twice; dvipdt, xiii.2.27Ms ; 3.25 (accent I cf. RV.) 
(itipd't, iii.29.1,2,6 ; sdrvapdt, x.10.27 ; sahasrapdt, vii.41.2. 

(6) The stem-vowel is unchanged. Forms (arranged by roots) : karambha-dt 
kravya-, 2 ; vifva-; havya-; — prathama-chdt ; — ukha-chit (see N.p.n.) ; — adri-bhit 
ltd-; gotra-; pur-, 5; — dkshetravit, 2; apva-vit, 2; kratu-, 4; kshetra-; gdtu-, 6 
go-, 3 ; draviifo- ; nabho- ; rayi-, 2 ; vaco- ; vayund- ; varivo-, 6 ; vasu-, 5 ; vicva- 
7 ; pruta- ; suar-, 5 ; hirayya- ; — adma-sdt, 2 ; antariksha- ; vpastha- ; rta- 
turaqya- ; durona- ; duvanya- ; prdgharma- ; vara- ; vioma- ; susarh- ■ cam&shdt 
dru-; wr-; vedi- ; pud-; — from AV., gdtu-vit, ndtha-, papa-, prajd-, vira-, xi.1.15 
purdtfa-; vasu-; sarva-; suar-; — trikakut, iv.9.8 ; udbhif, v.20.11; su-hd'rt, ii.1.5. 

Transition to a-declension : pd'da-s, p. 471. 
Feminine. There are 10 forms (from 9 stems). 

Forms: (a) apd't, 2; (6) kdkut; dhrshat; bhasdt; par at; sarhvit ; sarhsdt; 
suyavasd't; from AV., kakut, vi.86.3 : ix.4.8; 7.5: x.9.19; drshdt, ii.31.1 ; bhasdt, 
3 times ; pardt, 3 ; sudsdt, xvi.4.2 ; nishdt, xx.132.6,7 (ed. vani-). 

Quite isolated stands the form sadha-md's, N.s.m., vii.18.7 ; ef. 
p. 463. The regular form sadha-md'd, p. -md't, occurs iv.21.1. 

Nominative and Accusative Singular Neuter. 

(a) We have the compounds of -pad with lengthened vowel : 
dvipd't and cdtushpdt (the forms with d occur together, and those 
with d together also), iv.51.5 :. x.27.10 ; tripd't, x.90.3 ; from AV., 
dvipd't and cdtushpdt, vi.107.1-4 : viii.8.14 : xix.31.4 ; 34.1 ; 
tripd't, ix.10.19. 

(b) On the other hand we have : dvipd't and cdtushpdt, i.49.3 ; 
94.5 ; 124.1 : x.97.20 ; raghu-shydt (dnikam), iv.5.9 ; prakala-vit 
(? as adv.), vii.18.15. 

Accusative Singular Masculine and Feminine. 
Masculine. There are 58 forms (from 30 stems). 

(a) Forms: apd'dam, 2; tripd'dam; pd'dam; sahdsrapddam ; sadha-md'dam ; 
from AV., pd'dam, 3 ; piUpd'dam, iii.29.3,5. 

(6) Forms (arranged by roots): kravya-ddam, 2; vipva-; — ud-bhidam; gotra-, 
2 ; pur-, 2 ; — ahar-vidam ; kratu- ; kshetra- ; gdtu- ; go-, 3 ; varivo- ; vasu-, 6 : 
vipva-, 3; sacir ; suar-, 12; hotrd-; — garta-sddam ; vanar- ; susarh-, 2; apsu- 
shddam ; dru- ; dMr-, 3 ; barhi- ; — raghu-shyddam, 3 ; havana-syddam ; from 
A V., gdtu-vidam ; go-; suar-; trikakudam, v.23.9 ; durhd'rdam, viii.3.25. 

Transition to a-declension : pdflcapdda-m, p. 471. 
Feminine. There are 15 forms (from 14 stems). 

Forms: upasadam; kdkudam; drsh&dam ; nidarh-nidam ; nividam,2; pari- 
pddam, ; pitrshddam ; barhisbddam ; vacovidam ; vipvavidam ; pojrddam ; sam- 
vidam ; sarhsddam ; samddam ; from AV., dyur-ddd-am, vi.52.3 ; bhasddam ; 
govtdam ; samvidam ; parishddam ; samddam. 

Instrumental Singular Masculine, Feminine, and Neuter. 
Masculines : dpadushpadd; kshetravidd; cdtushpadd; dvipddd; 
padd', 12 ; varivovidd; suarvidd; from AV., padd', 3 ; sushddd, 



I.s.m.f.n.] Noun- Inflection in the Veda. 473 

iii.14.1 : ii.36.4 (read -dm?). Here BR. (vii.603) put sdd-d^ivA.7. 
The accent is irregular. 

Feminities : uda, 2 ; upavidd: drshddd (and AV., 2) ; nividd, 
2 ; nishddd, 2 ; pran'idd; mudd ; vidd' ; samsddd; suarvidd. 

Neuters: hrdd", 21 (and AV., 3); dushpddd ; dvipddd and 
cdtushpadd, AV. xix.15.2. 

Adverbial shift of accent is seen in sarvahrdd' , for -hr'dd, 
x.160.3. See p. 358 top. 

Dative Singular Masculine, Feminine, and Neuter. 

Masculines: apdde ; kuhacidvide ; kravyd'de; cdtushpade; 
vedishdde; sadandsdde; suarvide; from AV., tadvide; sarvavide; 
suarvide. 

Feminines: nid'e, 8; mud'e, 2 ; pitrshdde; parade, AV. viii.2.22. 

Neuters: hrdb, 14; dvipdde, 11; cdtushpade, 11 ; from AV., 
hrde, 3 ; cdtushpade, vi.59.1. 

In cdtushpdde ndridya dvipd'de, i.l21.3<?, the 4 of the penulti- 
mate is purely metrical ; see A.s.n. (a). 

Ablative Singular Masculine, Feminine, and Neuter. 

Masculines : dvipddas; padds, AV. ix.5.3. 
Feminines: uttdndpadas, 2; kdkudas; nidds, 11 ; sarhvtdas, 
AV. iii.5.5. For niddyds, see p. 4 VI. 

Neuters : hrdds, 6 ; from AV., hrdds, 3 ; shdtpadas, xiii.2.27. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines : nrshddas; yavd'das; suarvidas; from AV., 'eha- 
padas; suarv'tdas, 5 ; raghushyddas; durhd'rdas, 6. 

Feminine : saihsddas, AV. vii.12.3 ; pramudas, RV. x.10.12, is 
A.p.f. 

Neuters: dvipddas and cdtushpadas, 2 (and AV. iv.28.1,6) ; 
hrdds, vii.101.5 : viii.18.19 (Gr. — text htdds) ; hrdds, AV., 9. 

Locative Singular Masculine, Feminine, and Neuter. 

Masculines : pad'i, 2 ; suarvidi, x.88.1 : AV. xvii.13. 

Feminines: nishddi, 2; carddi ; samsddi, 2; kakudi, AV. 
iii.4.2: vii.76.3. 

Neuter: hrd'i, 13 ; AV, 12. In vi.53.6, hrdi may be an A.s.n. 
of stem hrdi (see BR.) ; and perhaps we have the same stem in 
x.91.13 and ii.23.16. 

Vocative Singular Masculine. 
Here belong : vicvavit, ix.64.7 ; sdrvavit, AV. vi.107.4. 

Nom., Acc, and Voc. Dual Masculine and Feminine. 

I. Masculines : (a) pd'dd, 6; (*) aharvidd, 2; kavichddd ; 
kratuvidd; vasuvidd; vicvavidd; suarvidd. 



474 C. R. Lanman, [Ead. d-stems. 

In AV. xviii.2.12, we have pathi-shdd-i, an ungrammatical 
imitation of pathi-rdksM, RV. x.14.11. 

Feminines : udbhidd; barhishddd; rapstidd ?. 

II. Masculines : (a) pd'dau ; from AV., pd'dau, 5 ; pddau, 
i.27.4; (b) dtmasddau,v. 9.8. Feminine: prdntasddau, AV. vii.95.2. 

Instrumental and Ablative Dual Masculine. 

Instr. : padbhyd'm, AV. v.30.13 : xii.1.28. Abl. : padbhyd'm, 
RV, 2. 

Genitive and Locative Dual Masculine and Feminine. 

Gen. masc. : padds, AV. xii.4.5. Loc. masc. : padds, RV. 
Loc. fem. : prdpados, AV. vi.24.2. In AV. i.18.2, padds, as well 
as hdstayos, does duty for an Abl. See pp. 344 med., 392 med. 

Nominative and Vocative Plural Masculine and Feminine. 
Masculine. There are 64 forms (from 33 stems). 

(a) Forms : apd'das ; cdtushpddas ; dvipd'das ; (itipd'das ; sadhamd'das, 8 ; 
somamd'das ; from AV., dvipd'das, xii.1.51; cdtushpddo dvipa'do yantt yd'mam, 
x.2.6 (<S metrical ?). 

(6) Forms (arranged by roots): Hrja-adas ; purusha-ddas ; madhu- ; yavasa-, 
2 ; soma- ; havir- ; — trdas ; pra- ; — deva-nidas ; — ud-bhidas, 4 ; — ahar-vidas ; 
ntthd- ; vaco-, 2 ; varivo-, 2 ; vasu- ; suar-. 1 ; hotrd- ; — adma-sadas ; pwah-, 2 ; 
farma-, 2 ; svddiisam- ; camA-shddas, 5 ; dhd/r- ; barhi-, 2 ; vanar-, 2 ; oarhi- 
shadas ; — raghu-shyadas, 4 ; vanadas (suffix ad) ; from AV., abhtmodamfrdas, 2 ; 
svddisammvdas, vii.60.4 ; as voc, xiii.1.3 ; — anna-vidas ; g&txir (voc.) ; ni- ; nt-; 
brdhma-, 12; yajur-; vifva-; simr-, 2; — anlariksha-s&das, 2; upa-, 2; divi-, 2; 
sahhd-, 3; — raghu-shyadas; — suhd'rdas, ili.28.5 : vi.120.3. 

Transition to «-declension. For pd'da-s and pd'dds, see p. 471. 
Feminine. There are 14 forms (from 10 stems). 

Forms: dmd'das; nidas; nishpudas; pramudas; mudas; vifvasuvidas; farddas, 
6 ; samsadas ; suhutd'das ; suarvidas ; from AV., pramudas, 2 ; upasddas, 2 ; 
agharitdas, 2 ; dwhd'rdas, xiv.2.29. 

Nominative and Accusative Plural Neuter. 

Roth suggested as a Vedic canon that compound stems in 
radical d or root -4- t make no plural in -ndi or -nti, but keep the 
form of the singular. Thus ukha-chU might be taken with pdrva, 
i. e. pdrvdni, in iv.19.9; so Say. Compare dirgha-priit (vratd'), 
viii.25.17 ; H'dhar (divid'ni), i.64.5. The parallelism of ii.15.7c, 
however, favors our taking it as JST.s.m. (cf. viii.68.2c). ' The 
(broken) decrepit one walked ; his members united together.' 
See BR. vii.1'714. The form hr'ndi is avoided by the use of 
hr'dayd, -dni. 

Accusative Plural Masculine and Feminine. 

Masculines : kravyd'das, 2 ; tudnldas; devanidas, 2 ; parishdd- 
as ; pastiasddas ; bdhukshddas ; from AV., cdtushpadas and 
dvipddas, 4 ; apsushddas, 2 ; sattrasddas; durhd'rdas, 11 (xix.28 
and 29) ; suhd'rdas, xix.32.6 ; padds, iv.15.14 : ix.4.14 : and RV. 
i.146.2 (never pddas). 



A.p.m.f.] Noun-Inflection in the Veda. 475 

Feminines : bhidas ; nividas, 3; pramtidas (x.10.12); vasu- 
vidas; vintidas; parddas, 19 ; samddas; sasyddas; havyasU'das, 
2 ; nidds, 4 (never nidas) ; from AV., parddas, 20. 

Instrumental Plural Masculine, Feminine, and Neuter. 

Masculines : gharmasddbhis, 2 ; patdpadbhis; padbhis, ' feet,' 
AV. iii.7.2: iv.11.10; 14.9: xix.6.2. 

By no phonetic process can pad-bhis become pad-bhis; but this 
is the regular resultant of pap-bhis (cf. vid-bhis, spat, vipdt). 
From pdp, ' glance' ( pap : spdp : : tar : star), we have padbKis, 
iv.2.12, dr'pydn padbhih papyer ddbhutdn: 'With thy glances 
beholdest thou the visible and the invisible.' 

Iniv.2.14; 38.3: v.64.7: x.79.2 ; 99.12: VS. xxiii. 13, however, 
we have padbhis evidently meaning 'feet.' The AV. has the 
regular forms with d ; cf. also padbhyd'm. Perhaps those with 
d are false forms due to false analogy or to confusion with padbhis, 
'glances;' but cf. pddgrbhi, x.49.5, and pddbtpa, i.162.14,16: 
x.97.16. 

Feminine : parddbhis, 2. Neuter : hrdbhis, 2 ; and AV. ix.1.1. 

Dative Plural Masculine. 

Here belong : prthivi-shddbhyas, AV. xviii.4.78 ; divi-, 80 ; 
antariksha-sddbhyas, 79. 

Genitive Plural Masculine and Feminine. 

Masculines : admasdddm; dvipdddm; from AV, cdtushpaddm 
and dvipdddm, 4; pvdpaddm, viii.5.11: xix.39.4 ; sushdddm, 
iii.22.6 ; suhd'rddm, iii.28.6. 

Feminines : paripdddm; samdddm; from AV, pardddm, xviii. 
4.70 (MSS. -dam) ; prdntasdddm, i.32.2. 

Locative Plural Masculine, Feminine, and Neuter 
Masculine: patsu, v.54.11 : AV. vi.92.1 ; -pdtsu!, Ait.Br. vi.2. 
Feminines : samdtsu, 32 times ; pardtsu, AV. xii.3.34,41 : xviii. 
2.38-45. Neuter: hrtsu, 9 times, and AV. viii.8.2 : xii.2.33. 

STEMS IN RADICAL DB. 

Here belong about fifty stems, from the roots idh, Jcshudh, nadh, 
bddh, budh, yudh, Xrudh, 2rudh, vidh (i. e. vyadh), sddh, sidh, 
sridh, ardh, mardh, vardh, spardh. The last four appear of 
course with the r-vowel. The verbal -vr'dh is the most common 
of all. 

There are no stems in which final dh is not radical." Gr. derives 
purudh from pardh, with svarabhatoi. The stem ishidh seems to 
be shortened from nish-shidh; cf. (n)i-shkrti. Agnl'dh is plainly 
shortened from agni'dh. The word prksMdhas is unclear. 
Budh suifers transfer of aspiration in ushar-bhUt. 

The forms are alike for m. and f. There are 7 monosyllabic 
fern, stems: hsMdh, nddh, mr'dh, yiidh, vr'dh, spr'dh, sridh. 



476 C. B. Lanman, [Ead. dh-atems. 

Masc. are vr'dham and vrdhd'm (adj.), and bd'dhas (subst.). 
Neuter forms (4 only) are found in the I.s. and G.s. 

No stem shows the distinction of strong and weak forms. 

Transition to the i-declension. The inf. yudhdye (accent — ef. 
yudh-i) is a transition-dative to y&dh, entirely equivalent to 
yudh-h in meaning, but not in metrical value. There is no stem 
yudhi. 

Irregularly accented are: sridhds, A.p., once (but sridhas, 15) ; 
nddbhyas. For vimr'dhas, see GLs.m. 

Nominative Singular Masculine and Feminine. 

Masculines : agni't, 2 ; antir&t; ushar-bhut, 2 ; yaviytit; pvd-vit, 
AV. v.13.9: VS. xxiii.56: xxiv.33. Here BR. place samitsamit, 
'flammend,' iii.4.1 ; but see I.s.f. 

Feminines : pravr't; samit, 3 ; from AV., samit, 3 times, as 
x.5.43 ; kshUt, ix.7.12 ; virtit, 8 times, as ii.8.2-4. 

Accusative Singular Masculine and Feminine. 

Masculines: vr'dham, 2; anna-vr'dham; dhuti-; gird-; tugrid-, 
2 ; namo-; payo-, 2 ; parvatd-; madhu-; vayo-; sadyo-/ saho-, 
2 ; su-, 2 ; — agnl'dham; asridham; wharbtidham, 3 ; goshu- 
ytidham; yajnasd'dham, 3; from AV., yqjnavr' dham, iv.23.3 ; 
marmdvidham, xi.10.26 ; hrdaydvidham, viii.6.18. 

Feminines : asridham ; usharbudham ; kshtidham, 3 ; nish- 
shidham ; y&dham ; virHdham ; samidham, 6 ; sridham ; from 
AV., samidham, 5 ; hshiidham, 2 ; ytidham, 2 ; virtidham, 2 ; 
sarhriidham, vii.50.5. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculine : suvr'dhd, ii.23.9. 

Feminines: yudhd', 19 and i.174.4; vrdhd' : samidhd, 23; 
sushamidhd, 2 ; from AV., samidhd, 5 ; Jcshudhd, iv.7.3 ; yudhd', 
i.24.1 : x.6.16 ; subMhd, xiv.2.31,75 (BR. as N.s.f. of stem -dha) ; 
suvr'dhd, ii.13.5 (BR., -d[s], as N.p.m. of stem -dha). 

In iii.4.1, samit-samit swmdnd bodhi asm'e,, pucd'-pucd sumatim 
rdsi vdsvah, BR. take the first word as N.s.m. So Say., atyar- 
tham samiddhas tvam. The parallelism of pucd'-pucd suggests 
the possibility of its being an Ls.f. without ending. Cf. vi.48.1aJ, 
and as-stems, I.s.n. 

Neuters: payovr'dhd; yaviytidhd; sdkamvr'dhd. 

Dative Singular Masculine and Feminine. 

Masculines : usharbtidhe; rtdvr'dhe; purunishshidhe ; mahi- 
vr'dhe; samr'dhe (personified — 'Bonus Eventus'?), AV. iii.10.10. 

Transition-form: yudhdye (inf., 7) ; see above. In i61.13c, we 
ought, perhaps, to read this longer grammatical form in place of 
its equivalent yudh-'e. 

Feminines: kshudhi; vrdh'e, 34; samidhe; yudhe, AV. iv.24.7. 



Ab.s.i] Noun- Inflection in the Veda. 477 

Ablative Singular Feminine. 
Here belong : kshudhds; yudhds; sridhds. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines: goshuyudhas; vi-mrdh-dsf,x,152.2: AV. viii.5.4,22. 

Anomalous accent. BR. see a G.s. in x.152.26, vrtrahd' 
vimrdhd va.pt. The only parallels that I can now adduce are 
parihvrtd', viii.47.6, and ddhid', AV. vi.132. For avadyabhiyd' , 
see p. 381 end ; the accent of sarvahrdd' , p. 473, is adverbial. 
Say. takes vimrdhd as N.s.m. — samgrdmakdri — and the vowel- 
stem occurs TS. ii.4.2', iydm vimrdhd' (tan&'s). But is it not 
possible that the original form of the verse was veretrahd' mrdhd 
vapi', and that the vi was slipped in from the other verses (vi 
mr'dhojahi, 3a, 4a) ? The forms of the other texts would then 
rest upon this error in the Rik. See Aufrecht, Jiigveda', preface, 
p. xlii and BR. vi.1143. 

Feminine : kshudhds. Neuter : hrdaydvidhas. 

Locative Singular Feminine. 
Here belong : prabudhi; mrdhl; yudhi, 4 ; sprdhi. 

Nom., Aoc, and Voo. Dual Masculine and Feminine. 

I. Masculines : asridhd, 2 ; rtdvrdhd, 7 ; rtdvr'dhd, 4 ; namo- 
vr'dhd; puroyddhd. 

Feminines : rdil-vr'dhd; ghrtd-; payo-; vayo-; sdkarh-. 

II. Masculine: rtdvr'dhan, i.23.5 ; as voc, i.2.8 : AV. iv.29.1. 
Feminine: samklhau, AV. xi.5.9. 

Nominative and Vocative Plural Masculine and Feminine. 

Masculines : rtd-vrdhas; rtd-vr'dhas, 12 ; tugrid-; payo-; par- 
vatd-; vayo-; suge-; su-, 2 ; — asridhas, 4 ; usharbddhas, 2 ; goshu- 
yudhas ; jnu-bd'dhas ; pray-ddhas ; vrshdyMhas ; purudhas ; 
somaparibd'dhas. > 

Feminines: amitrdyudhas ; asridhas; ishidhas ; rtdvr'dhas 
(dvd'ras), 2 ; nishshidhas, 3 ; paribd'dhas, 2 ; parispr'dhas ; 
mr'dhas; virudhas ; virddhas, 4; pirudhas, 2; samidhas, 2; 
spr'dhas, 2; from AV., kshddhas, xi.8.21 ; vtriidhas, 10 times; 
samidhas (N. and A.), 7. 

Accusative Plural Masculine and Feminine. 

Masculines : rtd-vr'dhas, 4 (and AV, 2) ; tamo-; rayi-;— 
usharbtidhas, 5 ; purtidhas, 2 ; bd'dhas, vi.11.1. 

Feminines : mr'dhas, 25 ; yudhas, 2 ; virudhas, 3 ; purudhas, 
5 ; samidhas, 3 ; sam-r'dhas; savr'dhas; spr'dhas, 21 ; sridhas 
15; sridhds (! ix.71.8) ; paribd'dhas (text' pari bd'dhas), viii' 
45.40: ix.105.6; from AV, y&dhas, x.10.24; sridhas (MSS. and 
ed. sr dhas), h.6.5 ; mr'dhas, 6 times ; virudhas, 7. 
vol. x. 64 



478 0. R. Lanman, [Radical 

Instrumental Plural Feminine. 
Here belong: samidbhis, 3; vir&dbhis, AV. v.28.5 : viii.7.15. 

Dative and Ablative Plural Feminine. 

Dative : n&dbhyas, at end of catalectic pdda x.60.6a — Gr., 
-bhias. BR. refer this to ndh (cf. akshdndhas, A.p.f.) ; Weber, 
Ind. Stud, xiii.109, to nap. Ablative : vir&dbhyas, AV. 

xix.35.4. 

Genitive Plural Masculine and Feminine. 

Masculines : rtdvr'dhdm; prabudhdm; vrdhd'm. 

Feminines : nishshidhdm / virudhdm, 2 ; sprdhd'm, 2 ; from 
AV., yudhdm {pate), vii.81.3 ; virudhdm, 13 times; virudhdm, 
iv.19.8. 

Locative Plural Feminine. 

Here belong : yutsu, 6 ; virutsu. 

STEMS IN RADICAL N. 

Under this category fall : go-shdn, tuvi-shvdn, svdn, tdn, dan, 
rdn, vdn, and 35 compounds of -hdn. The last are the only ones 
of importance. Almost all the forms are masculine. The stem 
tdn (tdnd, tand' tdne) is feminine. There is but one neuter form, 
dasyu-ghn-d' (mdnasd). The compounds of -gh(a)n form the fem. 
by adding l; thus, -glint. The fem. virahani (gadd) occurs MBh. 
ix.3238. BR. refer this to virahdn, s.v. ; but in vol. vii. col. 1513, 
perhaps better, to virahana. 

The distinction of strong and weak cases appears only in the 
compounds of han, and is here effected negatively by dropping 
the d in the weakest cases ; thus, pdd : pdd : : (g)hdn : ghn. The 
syncopated vowel is never to be restored for the sake of the 
metre, as is the case, e. g., with somapd'v(a)ne. 

The strong form seems to be used for the weak in Manu xi.101, 
ciravdsd dvijo , ranye cared brahmahano vratam ; so in 128. 
(Is the anomaly due to the metre? -ghno occurs viii.89.). So 
raksho-hdn-o, valaga-hdn-o, A.p., p. 481. 

In general, as we saw above, p. 442, the stems from roots end- 
ing in n go over to the vowel-declensions in d and &; thus from 
jan, prathama-ja'-s and -jd'-s. Most of the forms of -san and -tan 
belong to the vowel-declension (p. 438, 442). One might expect, 
d priori, forms like -jd', -jdnam, -jdn-d (-JM'), -jne, etc. ; but there 
is no evidence for such forms save in tbe stems mentioned at the 
beginning of the section. 

The Zend vSrethra-jdo seems -to be the reflex of a Sanskrit 
*vrtra-hd'-s; but no such form occurs, nor do the three instances 
of metrical hiatus warrant our inferring it (see N.s.m.). 

Transition to the a-declension. The N.A.s. neuter of stems in 
-hdn would properly be -h&'. This form seems to have been 
avoided and its place supplied by a form of transition to the 
a-declension. In the Rik we have : vrtrahdm (pdvas) ; satrdhdm 



n-stems.] Noun- Inflection in the Veda. 479 

(pavtnsiam) ; and the L.s., karanjah'e (vrtra-hdtye). In the AV. 
these transition-forms extend even to the masculine and we have : 
ardtihdm (paripd'nam — or n.), xix.35.2 ; patru-hd-s, i.29.5 : vi. 
98.3 ; sahasra-hd-s, viii.8.17. Further, we find the stems pdni-gha 
and tdda-gha, Pan. iii.2.55. 

The stem -ghna is, as I think, not an independent formation 
[gh(a)n-a], but rather a transition-stem starting from the weak 
forms of -han, which are sometimes ambiguous. Thus Say. sees 
in Ait. Br. viii.23, gangdydm vrtraghne ''badhndt paftca pancd- 
patar'n haydn, the Loc.s. of 'Vrtraghna, a district on the Ganges ;' 
but it is doubtless the D.s.m. of vrtrahan (BR.). In the Rik we 
have: dpHrushaghnas tndras, i.133.6; dhighne, vi.18.14; dpva- 
ghndsya, x.61.21 ; go- and pdrusha-ghndm, i.114.10; parnaya- 
ghne, x.48.8 ; sughnd'ya, viii.59.11 ; hastaghnds, vi.75. 14. Each 
form occurs only once. Cf. pdnighndm, VS. xxx.20. 

The -ghna-iovxas, in general, belong to a younger linguistic stratum. Thus we 
have rdksho-h&n in the samhitd ; but rahshoghnaif ca s&ktaih, Kauc. 126. Indra 
and Qiva are called respectively valavrtrahan and bhaganetrahan in the old Vana- 
parvan of the MBH. ; but valavrtraghna and bhaganetraghna in the Anucdsana- 
parvan. The -ghna-iorms are common in post-Vedic texts (so artha-, papu-, 
bhrAna-, yafo-, visha-, in Manu), although, of course, the -ftore-forms do not die 
out. Both stems stand side by side in Manu viii.89 ; cf. viii.311 with iv.208. 

Transitions also from the strong forms of -han are seen in the 
later texts. Thus, analogous to the A.s.m. vrtra-hdn-am are 
formed : the neuters ripu-hana-m (cakram) and paratejo-hana-m 
(tejas) ; and the masc. stem mahdndga-hana (epithet of Civa) ; 
see BR. vii.1513. 

The root may therefore appear in six forms at the end of a 
compound : namely, as ghan, han; gha, ha; ghna, and hana; 
thus, (witli gh preserved after j) abhog-ghan, pairu-han, pdni-gha,, 
patru-ha, pdni-ghna, ripu-hana. 

It is not unlikely that the otherwise unauthenticated stem tdnas 
is^ due to a merely formal parallelism in v. 70.4c, md' peshasd md' 
tdnasd. Cf. ushdso doshdsap ca and as-stems, Gr.s.n. The metre 
would be good with md' tdnd mdtd peshasd. 

The accent of the oblique cases of monosyllabic stems is regular 
in only two instances : tand ', x.93.12 ; vand'm, x.46.5 (see p. 353). 
Elsewhere we have: tdnd, 19; tdne, 7; rdne, rdnsu; vdnsu ; 
svdnif. Since tdnd is often used as an adverb, we may say that 
its accent has suffered an adverbial shift — here recessive ; cf. diva, 
adv. from the l.s. div-d'. 

Nominative Singular Masculine. 

Here belong 105 forms (from 23 stems) : d-vtrahd; -hd", with 
aghapansa-, 3 ; adrshta-, 2 ; abhimdti-; amitra-, 4 ; amiva-, 3 ; 
arupa-; apasti-, 5 ; asura-; ahi-; krshti-; go-; dasyu-, 5 ; nr-; 
puro-; mushti-, 3; raksho-, 8; vasar-; ' vrtra-, 55; parya-, 2; 
satrd-; sapatna-, 3 ; sapta-; from AV., dtirndma-; prpni-; mano-; 
mushti-; ydtu-; raksho-, 4 ; vrtra-, 9 ; sapatna-, 2l'; dvirahd, 2. 
See also TS. i.3.2. Scansion of the pddas in which these forms 



480 0. R. Lanman, [Bad. n-stems. 

occur shows that -hd stands before a vowel with metrical hiatus 
only in i.186.6 : ii.20.7 : x.74.6. 

For ghrdns, AV. vii.18.2, see rad. «-stems, N.s.m. 

Nominative and Accusative Singular Neuter. 

No example. Paradigm : vrtra-ha 1 . In place of this a form of 
transition to the a-declension is used (p. 478) : satrd-hdm, v.35.4 ; 
vrtra-hdm, vi. 48. 2 1 bis. 

Accusative Singular Masculine. 

Here belong : dprahanam; abhimdti-hdnam; ahi-, 2 ; tamo-; 
dasyu-; raksho-, 3 ; vrtra-, 7 ; satrd-; from AV., raksho-; memo-; 
sapatna-, ix.2.1. 

Instrumental Singular Masculine, Feminine, and Neuter. 
Masculine: vrtraghnd', 111 5.5: AV. iii.6.2. Neuter: dusyu- 
ghnd'. 

Feminine: tand', x.93.12; tand, 19 times. 

Tand is used as an adverb with recessive accent (p. 419) 10 times. With the 
game accent it is used : as adj., i.38.13 : ii.2.1 ; as subst., vi.49.13: vii.104.10,11 : 
x.148.1 : i.39.4. In i.26.6 : ix. 1 .6, I would take tand and (A(vatd both as adverbs 
separately. As a substantive tand is fern. ; but if we join (dfvatd (for -tyd t) with 
it, the anomaly has parallels: i.l22.11<2, pr&fastaye mahind' ralhavate (the metre 
favors the grammatically more correct -vatyai); ix.15.26, hrhatb devatdtaye; so 
ix.96.46. 

Dative Singular Masculine and Feminine. 

Masculines : asuraghne ; ahighne ; nrghne ; vrtraghne, 3 ; 
makhaghnk, TS. iii.2.4 1 " 3 ; bhrdnaghne, TBr. iii.9.15 3 ; further, 
rdne, RV. ix.66.13. 

Feminine : tdne (accent !), 7. 

Genitive Singular Masculine. 
Here belong: asuraghnds; vrtraghnds, 2, and AV. iv.24.1 : 
vi.82.1 ; further, go-shanas, RV. iv.32.22. 

Locative Singular Neuter. 

I. Here belong : tuvi-shvdni, 2 ; svdni, ix.66.9 (BR. as verb) ; 
bhrdnaghni, AV. vi.112.3 ; 113.2. 

II. The ending is dropped in : ran; ddn, 5. 

Vocative Singular Masculine. 
Here belong : amitrahan, 3 ; ahihan; vrtrahan, 32, and AV., 5. 

Nominative, Accusative, and Vocative Dual Masculine. 
I. Here belong : tamohdnd; rakshohdnd; vrtrahdnd, 2 (and 
AV., 2) ; vrtrahand, 4. 



N.V.p.m.] Noun- Inflection in the Veda. 481 

Nominative and Vocative Plural Masculine. 

Here belong : tuvishvanas; abhoggh&nas; punarhdnas; cutru- 
hdnas; matsyahdnas, Qat.Br. xiii.4.3 la ; yajnahdnas, TS. iii.5.4. 

Ace, Inst., Gen., and Loc. Plural Masculine. 

Accusative : sdtayhnds, AV. iii.15.5 ; raksho-hdn-o valaga- 
hdn-o vaishnavd'n (! p. 478), TS. i.3.2 quinquies. Instrumental : 
vrtrahdbhis, RV. vi.60.3. Genitive : vand'm (dhiyam dhus, 

aksharapankti ; Gr., vanddin), neut. ?, x.46.5/. Locatives: 
rdnsuj vdnsu, 2 ; cf. ddnsu. 

STEMS IN RADICAL P. 

Here belong the stems : (masc.) agnitdp, abhildpaldp, asutr'p, 
ketasdp, parirdp, pacutr'p, pras-dp, ritiap; (fem.) dp, dtdp, kr'p, 
kshdp, Jcship, patirip, rip, rdp, vishtdp; rtasdp and v\p (m.f.). 

The forms are alike for m. and f. There are no neuters. 

The distinction of strong and weak forms is seen in the stems : 
dp, rttiap; rtasdp, ketasdp. 

The weak form is used for the strong in apds (N.p.f.) ; but the 
texts show no small confusion in the forms of this word. 

The strong form is used for the weak in d'pas (A.p.f.) ; the d of 
parird'pas is metrical (see A.p.m.). 

Irregular accent. The character of hymn x.61 is a sufficient 
comment on the accent of vipas, verse 3, if it is G.s.m. Gr. takes 
kshdpas as G.s.f. twice ; but see G.s.f. The frequent A.p.f. of dp 
is almost always accented as a weak case, apds, perhaps for the 
sake of differentiation from dpas, ' work.' In like manner we have 
kshapds thrice, and vipds once (see A.p.f.). Twice, in the AV., 
even the N.p. is oxytone, apds ! 

Transitions to the a-declension. The ^>-stems do not show a 
single N.s. in the RV., nor a N.A. p.n. In the only places where 
these are necessary we find forms of transition : namely, ydtra 
bradhndsya vishtdpa-m (N.s.n.), ix.113.10 ; and imd'ni tri'ni 
vishtdpd (A.p.n.), viii.80.5. These are the only two forms in the 
RV. requiring the assumption of a vowel-stem vishtdpa, and they 
are plainly used to avoid the unfamiliar vishtdp (N.s.n.) and 
vishtdmpi, which general analogies would require. 

Almost all the later Vedic texts cited by BR. under vishtdpa 
contain the very form of the Rik-phrase bradhndsya vishtdpam; 
but Qat. Br. xii.3.1 9 has bradhnasya dste vishtape (L.s.). 

For the transition-forms kshapd'bhis and ' ksh'ipdbhis, see I.p.f. 

In the jagatt-pdda i.55.]c, perhaps dtapd-s is a transition-form, the extended 
stem being used rather than the N.s.m. dtdp, to fill out the otherwise short metre : 
so in iv.5.146, prati'tyena tcrdhund atrp&'sah. 

Nominative Singular Neuter. 

No example. Paradigm : sv-ap. In place of this a form of tran- 
sition to the a-declension is used : vishtdpa-m, by origin plainly 
the A.s.f. vishtdp-am used as a neuter nominative. See above. 



482 C. R. Lanman, [Rad. ^-sterns. 

Accusative Singular Masculine and Feminine. 

Masculine : papu-tr'parn. Feminine : mshtdpam, viii.32.3 : 
ix.34.5 ; 41.6 : and viii.58.7 (Gr., as a-form) : AV. xi.1.7 ; 3.50. 

Instrumental Singular Masculine and Feminine. 

Masculine : vipd'. Feminines : kshapd' ; vipd', 4 ; krpd', 8 ; 
apd', RV. viii.4.3. 

Ablative Singular Feminine. 
Here belong : dtdpas, 2 ; vishtdpas; apds, x. 95. 10. 

Genitive Singular Masculine and Feminine. 

Masculine : vipas !, x.61.3. Feminines : apds, 5 ; kshapds, 
2; ripds, 2; rupds, 3. If kshdpas, i.44.8 : ii.2.2, be G.s., 'by 
night,' we may consider the accent as adverbially recessive ; but 
in i.44.8, it is better taken as A.p.f., ' through the nights :' and in 
ii.2.2, as object of d' bhdsi. 

Locative Singular Feminine. 
The only example is vishtdpi — RV., 5 times and AV., 4 times. 

Nominative Dual Masculine. 

Here belong: ritidpd; asutr'pd u-, p. -au u-, x.14.12. The 
AV., xviii.2.13, has the more modern form and consequent sandhi, 
asutr'pdv w-. Cf. p. 341,4.a. 

Nominative and Vocative Plural Masculine and Feminine. 

Masculines : (a) rta-sd'pas, 6 ; keta-sd 'pas ; riti-dpas, voc. ; ( b) 
agnitdpas; asutrpas; prasHpas; vipas, 2 ; abhildpa-ldpas, AV. 
xi.8.25. 

Feminines: (a) d 'pas, 135; dpas, 17; rtasd'pas; from AV., 
d'pas, 94 ; d'pas, voc, 1 ; dpas, 27 ; ( b) patirtpas; ripas, 2 ; vipas; 
kshipas, 13. 

In i.70.7, 1 regard kshapds as A.p.f. ; see p. 422 med. The N.p. 
of dp is oxytone and weak, dpds !, in two Atharvan passages : 
ii.3.6 (ed. dpds), and vi.23.3. 

Nominative and Accusative Plural Neuter. 

No example. Paradigm : sv-dmpi or sv-dmpi. In place of 
*vishtampi, a transition-form is used : vlshtdpd; see p. 481. 

Accusative Plural Masculine and Feminine. 

Masculines : asutrpas; parird'pas, p. -rd'pah, ii.23.3,14. The 
d is metrically justified in verse 14, but not in verse 3. See Prat. 
ix.26. 

Feminines. I. The accent is that of a strong case (regular) : 
kshdpas, 6 times, and probably in i.44.8 and ii.2.2 (see G.s.f.) ; 
ripas, twice ; vipas, thrice. 



A.p.m.f.] Noun- Inflection in the Veda. 483 

The strong form d'pas is used for the weak : in books i. and x. 
of'theRV. (1.23.23; 63.8; 190. 7: x.4.5 ; 9.9; 121.8); intheAV., 
16 times, as i.25.1 : xiv.1.39; in the Ait. Br.,viii.l7 (dtapavarshyd 
dpo 'bhydniya). 

II. The accent is that of a weak case (irregular) : apds, 152 
times (and AV., 26) ; vipds (also vipas), viii.52.7 ; kshapds, 
iv.16.19 : viii.26.3 : and i.70.7 (p. 422 med.). 

Instrumental Plural Feminine. 

The forms falling under this category are particularly instructive 
as showing the working of linguistic tendencies, which, although 
unconscious, are none the less real. 

There is occasion in the Vedic texts for using the I.p. of the stems dp, samsr'p, 
kshdp, and kship. The organic forms would be abbhis, sariisr'bbhis, kshabbhis, and 
kshibbhis. 

I doubt whether the combination bbh, although given by Benfey (Gram. p. 23, 
1. 2 — I have not access to Bdhtlingk's Bemerkungen), ever really occurs in Sanskrit. 
Abbhra rests on a false etymology of abhrd ; cf . dbhri, abbhri. The group does 
not occur once in the AV.; see Whitney's additional note 3 to Ath.Pr. p. 588 
(256). If it occurs at all, it is, at any rate, of the extremest rarity. 

This unfamiliar combination, accordingly, is avoided outright by the Vedic 
language, and that in two ways : either (I.) by phonetic dissimilation of the two 
labials ; or (II.) by using a form of transition to the (^-declension. 

I. Here belong : adbhis, 15 (and AV., 5) ; sarhsr' dbhis, TBr. 
i.8.1 1 {tat saihsr'dbhir dnu sdm asarpat \ tat samsr'pdm samsrp- 
tvdm) ; cf. nddbhyas (Weber, stem ndp) ; kakhd, p. 471 end. 

Disregarding the Hindu systems of duplication in consonant groups, we see 
that the labial of the stem dp would become lost or obscured when pronounced 
with the labial of the ending (abbhis, abhis). The individuality of the word would 
thereby suffer greatly. To me it seems certain that the unconscious recognition 
of this fact motivated the phonetic differentiation of bbh to dbh. A similar motive 
has caused the retention in Sanskrit of the old Vedic ending in the I.p. of the 
pronoun d, ebhis ; if modernized to ais, it would scarcely be felt as an independent 
word. See p. 349 ad fin.; and Weber, Ind. Stud, xiii.109. 

II. Forms of transition to the (^-declension — kshapd'bhis, kship- 
dbhis — are found in the only passages where a consonantal (" mid- 
dle") case of the stems kshdp and kship is required. 

These are: iv.53.7c, sd nah kshapd'bhir dhabhif ca jinvatu; and ix.9'7.57c, 
hinvdnti dhi'rd dafdbhih kshipdbhih. These forms were undoubtedly made for the 
occasion. It is no explanation of them to set up the stems kshapd' and kshipd. 

The point of departure for the transition-form kshapd'bhis was probably some 
oxytone weak ease of kshdp : as, kshap-d'; and hence the accent (cf. ndv-d', I.s., 
with ndvdyd; vrt-d', Lb., with vrtdyd; ish-i, L.s., with ishdye). Prom kship, the 
paroxytone N.p. kshipas was the commonest form, and this may account for the 
accent kshipdbhis. Cf. id dbhis, sA-stems. 

Ablative Plural Feminine. 

Here belong: adbhyds, 5; adbhids, ii.1.1 : vi.62.6 : x.39.4 ; 
63.2 ; adbhyds, Gr. -bhids, i.34.6 and 112.5 (at end of pdda of 11 
or 12 syllables), i.80.2 (at end of pdda of 7 or 8). The AV. has 
the form adbhyds 5 times. 



484 G. R. Lanman, [Rad. bh-stems. 

GENITIVE PLURAL MASCULINE AND FEMININE. 

Masculine : vipd'm, 3. 

Feminities : kshapd'rn; vipd'rn,\x.<i<t).\; samsr'pdm, TBr. i.8.1 1 ; 
apd'rn, RV, 85 times (and AV., 48) ; dpdm (napdt), RV, 2 (and 
AV. vi.3.1,3) ; further, apd'm has the value of a spondee 8 times 
(cf. Kuhn, Beitrdge, iii.476). 

Gr. reads apddm, with resolution, in 16 passages ; but in 
x.46.1,2, we have, perhaps, aksharapankti. For 493.14, see Ueb. 
i.580. Bollensen, Z.D.M.G. xxii.586, reads apd'dm, i.67.10. 
Finally, the cadence requires dpddm or dpd'm at the end of 
i. 46.4a and viii.25.14a (7 or 8 syllables), and at the end of 
ix. 108.10c (11 or 12 syllables). 

Locative Plural Feminine. 

Here belongs aps-u — RV., 81 times, and AV., 43 times. BR., 
i.292, take apdsu, viii.4.14, for apsti,, with inserted a, as in dp-a- 
vdn, AV. xviii.4.24. Of. samsrp-a-ishti. See, however, as-stems, 
L.p.m.n. 

STEMS IN RADICAL BE. 

Here belong the stems: (masc.) jiva-gr'bh, sute-, sytima-; rathe- 
p&bh ; gharma-st'&bh, chandah-, vrsha-; sushMibh ; trikakiibh ; 
(fem.) kakHbh; ksh&bh; gr'bh; ddbh?; ndbh; gubh; anu-sht'&bh, 
tri-; rta-sttibh; sttibh and parishttibh (m.f.). 

The forms are alike for m. and f. There are no neuters. 

The distinction of strong and weak forms appears in ndbh. 1 
see no reason why nd'bhas (N.p.) and nd'bhas (A.p.) should not 
both be referred to ndbh. Perhaps kaktibh is a transition-stem 
from kakubhd; see p. 471. 

Nominative Singular Masculine and Feminine. 

Masculines : trikak-dp; st-kp. Feminines : trishtup, 2 ; from 
AV., kakup, xiii.1.15 ; unushtup and trishtup, viii.9.20 : xix.21.1. 

Accusative Singular Masculine and Feminine. 

Masculines: rathepubham, i.37.1 (see p. 330 ad init.) : v.56.9 ; 
sutegr'bham. 

Feminines : kakHbham; gr'bham; p&bham, 6 (and AV. xiv. 
1.32) ; rtasttibham; anu-shtubham ; tri-, 2 ; the last two, AV. 
viii.S.u! 

Instrumental Singular Masculine and Feminine. 
Masculines : sushtiibhd, 2 ; stubhd'. Feminines : anushtubhd; 
kshubhd'; grbhd'; 'fubhd', 2, and AV. xiii.1.21. 

Dative Singular Masculine and Feminine. 

Masculines : gharmasttibhe; sydmagr'bhe. Feminine : $ubh'e 
(as inf.), 17 ; AV. vii.106.1. 



Ab.s.m.£] Noun- Inflection in the Veda. 485 

Ablative Singular Masculine and Feminine. 

Masculine : jivagr'bhas. Feminine : trishtubhas, AV. xx. 
2.1-4. 

Genitive Singular Masculine and Feminine. 

Masculine: sushplbhas. Feminine: pubhdspdti, 5 (and AV. 
vi.3.3) ; pubhas pciti, 16 (and AV. vi.69.2 : ix.1.19). Were it not 
for the accent, we might take ddbhas as G.s.f. in v. 19.4, ddabdhah 
pdpvato ddbhah : ' undeceived by the deceit of every one,' 
' deceived by no man's deceit.' 

Nominative Plubal Masculine and Feminine. 

Masculines: chandah-stubhas, vrsha-; pari-shtubhas, su-. 
Feminines : (a) nd'bhas; (b) pubhas, 2 ; stubJias, 3. 

Accusative and Genitive Plural Feminine. 

Accusatives: kakubhas, 4; nd'bhas; stubhas; tri-shtubhas, 
pari-. Genitive : kak-dbhdm. 

STEMS IN BADICAL M. 

Here belong the stems : pdm, n. ; him, m. ; ddm, n. ; sam-ndm, 
f. ; kshdm {gdm &nAjdm), f. 

The first is indeclinable ; it occurs 84 times as N. or A.s. 

The stem him is posited for the sake of the I.s. himd', x.37.10; 
68.10 (cf. zim-d, G.s., Vend. i\.22bis) ; but this form can just as well 
be referred to the stem himd and might be added on p. 334.3. 

From ddm we have the form damd'm (Gr. -ddm), x.46.7 : VS. 
xxxiii.l (Mahldh., grhdndm). With this compare ddmpati, pdtir 
ddn, ddhsuf, rad. w-stems, L.s. and p. For d'e, v.41.1, a possible 
form of transition to the o-declension (L.s.n. — as though the A.s. 
were dd-m), see p. 449. 

From sam-ndm we have the N.p. samndmas, AV. iv.39. 1,3,5,7. 

The stem kshd'm has d in the strong cases. The forms are : 
kshamd', La, 5 (and AV. vi.57.3) ; kshmds, Ab.s. ; kshdm-i 
(accent !), L.s., 9 ; kshd'md, dual ; dyd'vd-kshd'md, 7 (and as voc. 
once) ; kshd'mas, N.p. Gr. proposes kshd'mam, text kshd'm, 
A.S., i.67.5 ; 174.7 : vi.6.4 : x.31.9 ; kshd'mas, text kshd's, A.p., 
iv.28.5 ; but see as-stems, A.s., A.p. 

Gr. refers gmds andjmds, Ab.s., to gmd' and jmd'; so kshmds 
might be referred to kshmd' (p. 448). 

STEMS IN BADICAL JR. 

Here belong the stems: (masc.) gir, tHr, m&r, vdrf; gdv-dpir, 
tri-, dddhi-, diir-, ydv-, sdm-; muhurgir; ajur, rta-, dhiyd-, sand-; 
ap-t-dr, dji-, rajas-, ratha-, radhra-, vrtra-, vipva-, supra-; sahdsra- 
dvdr; dur-dMr, su-; dmiir; apaspMr; ni-shttir; (fern.) gir, dvd'r, 
dh&r, pur, psur; dpir; amdjur, nijur; nipur, pardpur; abhipra- 
vol. x. 65 



486 C. B. Lanman, [Eadical 

m/far; upastir, vishtir, samstir; dnapasphur ; abhisvdr; (neuter) 
vd'r, s&ar; prtsutur, vipvatur, rdsdp/r; mithastur and sarkgir (m. 
andf.). 

The ar of dhar, tl'dhar, and vddhar appears to be suffixal. For 
vandhur and yanttir, see below. 

Only in vd'r, dvd'r, abhisvdr, and suar is the radical r preceded 
by any other vowel than i or w. The use of roots in ar as stems 
seems to be avoided in several ways : the root appears with r-\-t 
{-vr't, -bhr't) ; or else with suffixal a (vdra, bhdra, cdrcara) ; or a 
thematic i is added and the root-vowel often dropped (cdkri, 
jdghri, pdpri, pdpuri, bdbhri, sdsri) ; or, again, the stem goes into 
the a-declension by dropping the r (pra-hvd-s from hoar; cf. upa- 
hvar-d). But if the radical vowel takes some other coloring, the 
r remains as stem-final (cf. ratham-tard and ratha-tur). 

The stem dhar is supplemented by dhan, and this again by 
forms of transition to the a-declension. The stem ii'dhar is like- 
wise supplemented by iX'dhan, and some forms again must be 
referred to H'dhas. 

The forms are alike for all genders, except N.A. d. and p. n. 

The distinction of strong and weak forms is seen in the stem 
dvdr — strong dvd'r, weak du-r. 

The weak form duras, N.p.f., is used in place of the strong 
dvd'ras, i 188.5 ; and the strong form dvd'ras, A. p.f., is used in 
place of the weak duras, i. 130.3. 

Transitions to the a-declension. Starting from the strong cases 
dvd'r-am etc., come the forms dvd'ra-m dvd'rena (Cat. Br. iv. 
6.1''"), dual dvd're, dvd'rdni. See also N.A.s.n. The transition- 
forms pd'da-s etc. begin to appear in the samhitd; but those ol 
dvd'ra- occur first in texts of the Brahmana sort. Starting from 
the weak cases dur-as etc., come the forms patd-durasya, x.99.3 ; 
-eshu, i.51.3. In the later language the stem dvdra replaces the 
Vedic dvd'r; and in like manner pura, the Vedic pur. The mas- 
culine stem pd'd- remains masculine (pd'da-) ; but the feminines 
dvd'r- and pur- produce neuters (dvdra-m, pura-m). Su-dhura-s 
(N.s.m.) is a transition-form corresponding to sudhtir-am; so 
apratidhura-s, Cat. Br. xiii.4.2 1 " 2 . Here belongs, perhaps, dn-apa- 
sphurdm, A.s.f, vi.48.11. As is known, vd'r in the later language 
goes over to the «'-declension, vd'ri. 

On the other hand, vandhtir-as (N.p.m.) and bdndhur-d (I.s.) 
are probably forms of transition to the consonant-declension, the 
vowel-stem being here the older ; cf. Jcahubh, p. 471. 

There is no real justification for a stem yantur. It is supported only by 
yanturam, occurring twice. The origin of the form is as curious as certain. In 
iii.27.11a, agnim yanturam apturam, we have, I think, a brilliant example of the 
working of the tendency to formal parallelism, yan-td'r-am (nomen agentis of 
yam) being thus blindly accommodated to ap-tur-am (in which tur=tar is a 
radical). In viii.19.2J, agnim tdishva yant&'ram, the cadence has asserted its 
right to a short in the penultimate at the expense of the proper form (yant&'ram) 

Irregular accent : durds, A. p.f., ii.2.7. For the accent of td'r-d, 
m'r-as, and sdr-i, see G.s.n. 



r-stems.] Noun- Inflection in the Veda. 487 

The vowel preceding r is lengthened when r is or was followed 
by a consonant : thus, gi'r (for gir-s), girbhis. 

Nominative Singular Masculine and Feminine. 

Masculines: gi'r, x.99.11 ; muhurgi'r; durdpir; rajas-tu'r, 2; 
ratha-, vipva-, supra-tti'r. Here Gr. puts vd'r (i. 132.3) and vdar 
(x.93.3). The passages are corrupt. 

Transition-forms. For su-dhura-s, iii.38.1, see p. 486. For 
pra-hva-s (tishthan), Ait.Br. iii.9, see p. 486. 

Feminines: amdju'r; gi'r, 19; dhu'r, 3; pu'r, 5; from A V., 
dvd'r, ix.3.22 ; pu'r, x.2.31 ; dfft'r, ii.29.3 : TS. iii.2.8*. 

Nominative and Accusative Singular Neuter. 

Here belong: dhar, 6 and i.71.2 ; dhar-ahar, 6 times ; tt'dhar, 
26 (for i.64.5, see A.p.n. ; and for x.61.9, L.s.n.) ; vddhar, 11 ; vd'r, 
vdar, 11 (and AV. hi. 13.3: iv.7.1 : x. 4.3,4: xx.135.12) ; suar, 90. 
In the AV., suar has no other declensional forms. Cf. Rik Pr. 
i.31,32: iv.13; and Weber, Beitrdge, iii.385. 

Transition to the a-declension. The language avoids the use of 
a r-stem at the end of a compound in the N.A.s. neuter. Thus 
we have in AV. x.8.43, punddrikam. ndva-dvdra-m, not ndvadvdr; 
and in Cat. Br. iii.3.3 18 , sa-dpira-m {etad). 

Accusative Singular Masculine and Feminine. 

Masculines : turam, v.82.1 ; ap-turam, 5 times ; dji-j rajas-, 3 ; 
ratha-; vrtra-, 4 ; — ajuram; apasphuram; rtajuram; gdvdpir- 
am, 4 ; t/dvdpiram, 2 ; sahdsra-dvdram; sudhuram. BR. take 
turam, iv.38.7, as an absolutive of tur, used adverbially. For 
yanturam, 2, see p. 486. 

Feminines: dpiram, 13; upastiram ; g'vram, 7 ; dhuram, 5; 
pAram, 1 1 ; samgiram; from AV, dvd'ram (d), xiv.1.63 ; dhuram, 
v.17.18 ; puram, 26 times. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculine: bdndhur-d (see p. 486), AV. iii.9.4. 

Feminines: abhipramurd ; abhisvdrd, 2; dplrd ; gird', 67; 
gird'-gird; dhurd' ; purdl. 

Neuter : vipvaturd. In vi.49.3, str'bhir anyd' pipi$& su'ro 
anyd', p. su' rah, the demands of the parallelism and sense are 
well met by Grassmann's emendation, sU'r-d, l.s.n. ; but in the 
ninth place we need a short d. Possibly this was the very reason 
why the padakdra wrote su'rah. Cf. Beitrdge, iv.204. 

Dative Singular Masculine, Feminine, and Neuter. 

Masculines : gir'e; nishture. Feminine : upastire, 3. 
Neuter : s-dr-'e, iv.3.8 (for accent, see G.s.n.). For su're, viii. 
61.17, see G.s.n. 



488 G. B. Lanman, [Bad. r-stems. 

Ablative Singular Feminine and Neuter. 

Peminines: dhurds, x. 102. 10; nijdras. 

Neuter. Gr., Web., takes -d'dhar, i.64.5, as Ab.s. See A.p.n. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines : gdvdpiras; ydvdpiras; radhratdras. Feminine : 
amdj'dras. Neuters: rdsdpiras; sd'r-as,\1. For sd'r-ah, vi.49.3, 
see Ls n. In viii.61.17, tiditd sU'ra d'dade } p. sd're, the paddkdra 
is doubtless at fault ; read s-d'rah. 

In i.66.10 ; 69.10, sdar appears as an aptote in a genitive relation. 

Since the stem of s4'r-as is really a dissyllable {s'dar), the par- 
oxytone accent is perfectly regular ; but once we have sdr-'e, as if 
from a monosyllabic stem. See p. 408 med. 

Locative Singular Feminine and Neuter. 

Feminines: puri, 2 ; dhuri, 13, and AV. v. 17.15. 

Neuters. Stiar is used as a L.s. in i.52.9 ; 70.8 : v. 44.2 : x.20.2 ; 
61.14. In x.61.9, d'dhar is perhaps a L.s., 'nubilo.' If we regard 
the first part of dhar-divi as representing a case, it might be 
called a L.s. 

Nominative and Accusative Dual Masculine and Feminine. 

I. Masculines : vrtratdrd; sandjdrd; sudhdrd, 3. Feminines : 
dvd'rd, 7 ; mithasttird. 

II. Feminines : dvd'rau, 2 ; dhdrau, 2. 

Nominative and Vocative Plural Masculine and Feminine. 

Masculines : apturas, 4 ; dmuras, 2 ; gdvdpiras, 3 ; giras; giras, 
3 ; tridpiras; dddhidpiras, 7 ; durdhdras; dhiydjuras; nishtdras; 
rrvdras; vandhdras; bandhdras, AV. iii.9.3. 

Feminines : dvdras, 3 ; dvd'ras, 5 (and AV. v.27.7) ; dnapa- 
sphuras; amdjuras; pdras, 2 ; mithasfdras ; giras, 64 (and AV., 
3) ; giras, AV. i.15.2. In i.188.5, we have the weak form dtiras; 
see p. 486. 

Nominative and Accusative Plural Neuter. 

In i.64.5, duhdnti d'dhar divid'ni dhd'tayah, d'dhar appears to 
be an A.p.n. See p. 474, and cf. Ab.s.n. (r-stems). 

Accusative Plural Masculine and Feminine. 

Masculines: dmdras, 2; gdvdpiras ; giras, 2; mithastdras ; 
ydvdpiras; sarhgiras; sudhdras. 

Feminines : giras, 80 ; ddras, 24 ; dhuras, 2 ; pdras, 44 ; 
psdras; vishtiras, 2 ; sarhstiras; from AV., giras, 2 ; pdras, 3 ; 
nipdras, pardpdras, xviii.2.28. 

The strong form dvd'ras is used as an A. p., i.130.3. 

Irregular accent is seen in durds, ii.2.7. In i.41.3, purds is an 
adverb ; in vii.21.4, Aufrecht reads pdras; cf. BR., s.v. vi han. 

In x.99.11, Roth proposes sut[d]vd ydd yajatd diddyad gi'h, 
and takes gi'h as A.p.f. : ' Als der sehr starke (su-tdvds) Gott die 
Loblieder vernahm.' 



Lp.m.f.n.] Noun-Inflection in the Veda. 489 

Instrumental Plubal Masculine, Feminine, and Neuter. 

Masculine : ratha-tu"rbhis. Feminities : lptirbhis; 2pHrbhis, 
5 ; girbhis, 85 (and AV., 4). Neuter : dhobhis, 2. 

Dative Plural Neuter. 
Here belongs vdrbhyds, VS. xxii.25. 

Genitive Plural Masculine and Feminine. 

Masculine : sdmdpirdrn. Feminines : gird'm, 2 ; purdm; 

purd'm, 1 ; girddm, vi.24.1 ; purddm, viii.17.14 ; purddm, x.46.5 
(or -d'm, aksharapafikti). 

Locative Plubal Masculine, Feminine, and Neuter. 

Masculine : tdrshu. Feminines : girsJvd ; dhdrsh'd, 1 1 ; 

pd/rshd. Neuter: prtsutd' rshu. 

STEMS IN RADICAL V. 

The stem div supplements dM and did. The forms are : div-d', 
-'e, -ds, -i; dhardivi; pradiv-4, -as, -i; sudivas; see p. 432. 

From the stem di'v, f., 'play,' we have divk and divi. The 
stem dytiJ furnishes supplementary forms : dydvam, AV. vii.50.9 ; 
dyuve {-=.divt), vii.109.5. Cf. Ekady-A's and Kamadytivam (pp. 
406-7). 

STEMS IN RADICAL Q. 

Here belong the stems : (masc.) i-dr'p, etd-, td-, yd-, M-, ahar-, 
ddre-, yaksha-, suar-, sa-, sit-/ susadre; tveshd-samdrp, bhimd-, 
hiranya-; smamdr'p; upari-spr'p, divi-, ni-, mandini-, ratha-, 
hrdi-; vipvapic; supip; upapr'p; ddrdddip; dnarvip; prdtiprdp; 
dddd'p; purodd'p; %'p; spdp; vishpdp; (fern.) die; dd'p; dtfe; 
ndp; pap; prd'p: v'vpdp; v'tp; vrip; ddip; pracfip; pukrapip; 
upadr'p; samdrp; sudr' pikasamdrp ; dnapasprp; (m. and f.) 
mitkddr 1 c ; pipdnga- and ranvd-sarhdrp ; upa- and rta-spr'p; 
(neuters) mandndp; divispr'p; sddanaspr'p; ddredr'p; etddr'p; 
smamdr'p. 

The forms are alike for all genders, except in the N.A. d. and 
p. n., of which there is no example. 

The distinction of strong and weak cases is seen only in the 
nasalized forms of the N.s.m., q.v. I can cite no occurrence of 
forms like purodobhydm from a weak stem purodds. 

Transition to the a-declension. The declension of puro-dd'c is 
commonly supplemented by transition-forms in the later texts. 
The only Rik-forms, puro-dd's and puro-dd' p-am, belong to the 
p-stem. But later we have : puro-ddpa-s, Ait. Br. ii.9 ; -ddpena, 
VS. xix.85 ; -ddpdn, (^at.Br. iv.2.5" ; -dd' pa-vatsd, AV. iii.4.35. 
The Atharvan forms -dd'pau, ix.6.12 : x.9.25, may be referred to 
either stem. 

As forms of transition are probably to be regarded : tdrpam 
{yad), Manu i.45 ; tddrpam (enas) yddrpam, v.34; etddrpdni, etc. 



490 C. R. Lanman, [Rad. j-stems. 

The references given by BR. for tdrpa, etddrpa, tddrpa, yddrpa, 
sadrpa, are chiefly post-Vedie ; those for idr'p etc., chiefly Vedic 
Both vowel and consonant-stems occur side by side : thus, yd- 
drpdya .... tddrn, Qat.Br. i.3.5"; yddrpdt . . . tddrnn, vii.4.1'. 
Some forms are referable to either stem : thus, sadr'pau. 

Transition to the *-declension. The inf. drpdye (accent — cf. 
drp-i) is a transition-dative to dr'p, equivalent to drp-i in meaning, 
but not in metrical value. There is no veal stem drpi. 

There is no irregularity of accent. The L.s.f. prd'pi is paroxy- 
tone because equal to pra+ dpi. 

The forms show inconsistency in the treatment of final p (=&') 
similar to that seen in the verb (e. g. prdnak, abhi nat) : thus, spdt, 
dik; vidbhyds, vikshti. 

Nominative Singular Masculine and Feminine. 

1. Masculine. 1. From the nasalized form of the root drp 
come: (a) sadr'nk s-, TS. ii.2.8" ; (b) kidr'nn i-, RV. x.108.3; 
sadr'nn a-, i.94.7 : viii.11.8; 43.21; (c) idr'n, any 'ddr 'n, sadr'ti, 
prdiisadrn, VS. xvii.81 (cf. TS. iv.6.5 s ). See pp. 456 and 463, and 
Ind. Stud, iv.305 note. 

2. The form commonly ends in k. Here belong: upapr'k; 
etddr'k (x.27.24) ; tddr'k; divispr'k; nispr'k; yddr'k; ranvdsam- 
drk; suardr'k; hiranyasamdrk; hrdispr'k; idr'k, AV. iv.27.6. 

Feminines : upadr'k, 2 ; ndk ; ranvdsamdrk ; samdr'k, 4 ; 
mdr'pikasamdrk; from AV., dik, iii.27.1-6 ; dnapasprk, xiii.1.27. 
IL Masculines : vishpdt; spat, 3. 
Feminines : vipdt; vit, 3. 
III. The form purodd'a occurs twice. Cf. p. 463. 

Nominative and Accusative Singular Neuter. 

Here belong : mand-ndk f; etddr'k; susamdr'k ; sadr'k, TBr. 
i.2.6\ In RV. v.44.6, yddr'k and tddr'k may be neuters (adv.). 

Accusative Singular Masculine and Feminine. 

Masculines : upari-spr'pam; divi-, 7 ; hrdi-; — tveshd-samdrp- 
am; pipdnga-; ranvd-; — susamdr' pam, 3 ; — dflre-dr'pam; «w-, 
2 ; suar-, 4 ; — durd-ddipam; purodd'pam, 19 ; spdpam. 

Feminines : ddipam, 2 ; rtaspr'pam; dipam; pradipam, 2, and 
i.95.3 : iv.29.3 (text prd dipam) ; vipdpam; vipam; vipam-vipam, 
4 ; pukrapipam; samdr'pam, 2 ; from AV., pradipam; pipdnga- 
mmdrpam; dipam, 19 ; prd'pam, 7 times, as ii.27.1. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculines : vipvapipd; susamdr' pd; ipd', VS. xLl. Femin- 
ines : dipd', 2 ; pipd' ; pradipd, 4 ; vipd', 7 ; ddpd', i.127.7. This 
last BR. take as homophonous Ls.m. of ddpd'; but see Beitrdge, 
.463. Neuters: divispr'pd; d&redr'pd. 



D.s.m.f.] Noun- Inflection in the Veda. 491 

Dative Singular Masculine and Feminine. 

Masculines : dnarvipe; idr'pe, 4 ; divispr'pe; dtiredr'pe; from 
AV., tdr'pe, 3 ; dUdd'pe, i.13.1. 

Feminines : vip'e, 8 ; vipe-vipe, 5 ; from AV., samdfpe; dipt, 6 ; 
vipk 

Transition-form : drpdye (inf., 8) is dat. to stem dr'p; see p. 490. 

Ablative Singular Masculine and Feminine. 

Masculine : suardr'pas. Feminines : vipds, 2 ; samdr'pas ; 
from AV., samdr'pas, 2 ; dipds, 55 ; dipS-dipas, 5. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines: divispr'pas; sudr'pas, 2; suardr'pas, 2; hiranya- 
samdrpas; hrdispr'pas; from AV., prdtiprdpas, ii.27.1-6; upa- 
spr'pas, xx. 11! 7.2. 

Feminines : vipds, 4 ; vipds-vipas, 2 ; vipas {pate) ; dipds, AV. 
vi.98.3 : xv.5.1-66**. Neuter: sddanaspr' pas. 

Locative Singular Masculine and Feminine. 

Masculine: divispr'pi. Feminines: drpi; pradipi, 2 ; vipdpi; 
vipi, 3 ; samdr'pi, 5 ; from AV., prd'pi (ed. pr&pi), ii.27.7 ; dipt, 
25 times ; pradipi, 4 ; vipi, 2 ; samdr'pi. 

Vocative Singular Masculine. 
Here belongs tveshasamdrk. 

Nom., Ace, and Too. Dual Masculine and Feminine. 

I. Masculines: rtasprpd; divisprpd; divispr' pd, 2 ; mithtidr'pd; 
suardr'pd. Feminine: mithudr'pd. 

II. Feminine: vipau, ix.70.4. 

Nominative Plural Masculine and Feminine. 

Masculines : rta-spr'pas, 2 ; divi-; mandini-; ratha-; hrdi-; — 
dUre-dr' pas, 2 ; yaksha-; su-; susarh-, 2 ; suar-, 3 ; — tves'hdsam- 
drpas, 2 ; vipvapipas; supipas; stisadrpas; spdpas, 7 (and AV. 
iv.16.4) ; uparispr' pas, AV. v.3.10. 

Feminines: ddipas; upaspr'pas ; dipas, 3; pradipas, 8; vipas, 
35 ; samdr'pas ; from AV, dipas, 15 ; pradipas, 28 ; vipas (N. 
and A.), 18. 

NOMINATIVE AND ACCUSATIVE PLURAL NEUTER. 

No examples. Paradigm : grhavinpi. Instead of this a tran- 
sition-form is used : etddrpdni [duhkhdni), MBh. iii.579. 

Accusative Plural Masculine and Feminine. 

Masculines: ahardr'pas; bhtmdsamdrpas; spdpas, 3; suar- 
dr'pas; Kiranyasamdrpas. Feminines: ddipas; dipas, 7; pra- 
dipas, 3; vipas, 32; vripas ; samdr'pas, 2; from AV, dipas, 18; 
pradipas, 13; vipas (N. and A.), 18. 



492 0. R. Lanman, [Radical 

INSTRUMENTAL PLURAL MASCULINE AND FEMININE. 

Masculine: susarhdr'gbhis. Feminines: vidbhis ; padbhis, 
' with glances,' iv.2.12 (see page 475). 

Ablative, Genitive, and Locative Plural Feminine. 

Ablatives: from AV., vidbhyds, iii.3.3 ; digbhyds, iv.40.8: 
x.5.28: xiii.4.34. Genitives: ddipdm; dipdm (pate); vipd'm, 
39 ; vipddm, vii.9.2 : x.46.6 ? ; from AV., dipd'm, 5 times ; 
vipd'm, 1; vipdm, iii.20.2. Locatives: vikshti, 48 ; from AV., 
dikshii, 9; viksht'i, ii.2.1 : ix.5.19. 



STEMS IN RADICAL Sff AND & 

The character of the final sibilant of the radical stems in sh 
and s is in most cases essentially the same. It has therefore 
seemed advisable not to separate them. Usually the difference 
is merely one oi phonetic surroundings. All the stems ending in 
s preceded by a or d are put down as s-stems (category B) ; but 
those in which any other vowel or k precedes are entered with 
the sA-stems (category A). 

The derivatives of pds, with weakened vowel i (dpis, sudpis, 
pro-, vi-, sarh-pis), and bhi's, and dprdyus are put down as s-stems 
by BR. and Gr. 

A. The sA-stems are: (masc.) mti'sh and mdksh (or f.) ; dksh, 
in an-dhsh; aram-ishf , gavish, papuish; brhad-uksh, sdkamuksh; 
vdta-tvish ; anrta-dvish, asaca-, rshi-, edhamdna-, jara-, brakma-, 
dvidvish ; vipvd-p&sh, vipvd- ; ghrta-prush, pari- ; yajna-m/dsh ; 
prnga-vr'sh ; (fem.) ish, ush, tvish, dvish, prksh, risk ; nemann-lsh, 
sam- ; sdkam-uksh ; vy-ush ; sarhodksh; patidvish ; dnddhrsh, 
sudhr'sh ; supr'ksh ; abhra-prush, ghrta-, vi-; nimish, dnimish; 
eamrish ; prdvr'sh ; abhi-prish, doshani-, hrdaya-. 

B. The s-stems are: (masc.) hrtsu-ds ; an-d's, su-d's; jfld's; 
su-dd's ; anu'rdhva-bhds, dure-bhds, su-bhd's ; md's, su'ryd-md's ; 
vds ; a-pds, uktha-pds (root pans) ; stisampds, su-dpis (root pds) ; 
dprdyus ; (fem.) d-jnds, kdd's, nds; -pds, abhi-, ava-, nih-, pard- 
(root pans) ; pd's, d-pis, pro-, vi-, sam- (root pds) ; vi-srds, su-, 
svayarh- ; upds (unclear). Add candrdmds, m. 

The neuters are few : pr'esh, ghrta-pnish, dadhr'sh ? ; and d's, 
bhd's, dds, yds, parhyds. 

The sibilant is sometimes a mere root-determinative : compare uk-sh with ug-r&, 
6j-as, aug-mentum ; pf'k-sh withpr'c; prush within*; prdyus and pra + yu. 

BR., s.v. dd'svant, hesitate to set up a noun dd's, and do not venture to assume 
that s is a formative element. In the case of bhd's and bhd'svant, it is true, we 
may identify the s with that of the secondary root bhds ; and so with bhi's : but 
how shall we dispose of jnd's, d's (for tins according to Fick, Vgl. Wb. z i.268), 
and su-dd's f Md's, from md, is still more peculiar and is perhaps the relic of 
some old participial formation (mddbhis). Whatever be the character of the final 
s in these words, there is no question that this is the best place to enumerate 
them. 



sh, s-stems.] Noun- Inflection in the Veda. 493 

The distinction of strong and weak cases is seen in the stems : 
puns, pumdns ; nd's, nds; uktha-p&'s (cf. a-p&'s), -pd's. The 
pada has in all cases uktha-p&'s-. The d may indeed be metrical 
in x.107.6 : ii.39.1 : iv.2.16 (penultimate of a trishtubh), and 
vii.19.9 (in the eighth place) ; but the metre actually makes 
against the d in x.82.7 (seventh place). The long vowel is 
inherent in d's, jUd's, md's, a-yd's, etc. ; and so in kd's, susampds, 
and mit'sh (but cf. yajna-mii, 'sh-as, N.p.m.). 

Examples of consonamtal cases are very rare in the sanhitas. 
We have andk, dadhr'k, edhamdnadv'tt, viprut, vipriidbhis, punsu, 
and a few nominatives s. in -s / and from later texts, viprtidbhyas, 
prdvrt, dor-bhydm, dor-bhis, doh-shu, mds-sii. 

Supplementary themes. In mdd-bhis, -bhyds, we have a supple- 
mentary stem mdnt ; but cf. Schleicher, Comp. 1 p. 565, §260. 
Some think the two stems md's (md'ns, cf. ^r)v — stem pctjv?) 
and md(n)t phonetically reducible to the same original, like the 
forms of the suffix vans, va(n)t; but see Zeitsch. xxiv.70. 

The stem d6s is supplemented by doshdn as follows: sing., d6s 
(RV. v.61.5); doshnd (Raja-tar. iv.481) ; doshnds (Qat. Br. iii. 
8.3"); doshani- (AV. vi.9.2); dual, doshl (Kauc. 45) or doshdni 
(AV. ix.7.7:' Ait. Br. ii.6) ; dorbhydm (MBh. i.153); plural, 
dorbhis (Malav. 77) ; dohshu (Bhag. P. i. 15. 16). 

The defective stem isfi (id) is supplemented by idd (ird). 

idd, 11 ish-aa, 9 

ish-am, 75 iddm, 9 ish-a*. 63 

ish-d', 23 irf-4', 3 idayd, 5 tfdbhis, 6 

ish-e, 15 
ish-ds, 11 id-ds, 11 iddyds, 5 Uh-d'm, 8 iddn&m, 1 

ide Iddsu, 2. 

The stem ish is never used for a consonantal case. The occur- 
rence of sh at the end of a form or before * or bh is avoided by 
using the d-stem; thus, ide (yoc), idd, iddsu, tddbhls. I hardly 
venture to call these transition-forms ; although the fact that 
the great majority of the forms of ish (147) accent the theme 
would account for the accent of the stem idd (cf. p. 483). The 
relation between Jcshapds kshapd'bhis, kshipas kshipdbhis, and 
ishas iddbhis is one of close analogy. Perhaps the forms iddm, 
idayd, iddyds, and iddndm are improper extensions of the <2-'stem 
to cases where it does not rightfully belong. 

Transitions to the o-declension. These occur from the stems 
md's, dpds, nds, kd's, nimish, ish, and d's. 

The language-users formed by false analogy a N.s.m. pd'da-s to 
correspond to pd'd-am. In like manner they made for md's-am, as 
though it were md'sa-m, a N.s. md'sa-s, x.85.5d (Sdrya's wedding 
— see N.s.m.) ; so also a N.p.m. md'sds, iii.32.9c<? and yi.24.1ab 
(the passages look like modern reminiscences of old songs — cf. 
Aufrecht, Rigveda 1 , preface, p. xii) : vi.38.4: vii.61.4: x.89.13; 
and also md'sdn in two very late passages — iii.31.9 (an unclear 
and tumid hymn) and v. 78. 9 (verses 7-9 are a late fragment as 
the contents and the use of sarvdtah show). 

vol. i. 66 



494 G. R. Lawman, [Rad. sh, s-stems. 

The stem dpds, f., 'hope,' makes in the later language a com- 
plete transition to the (^-declension. In the Rik, all the forms 
come from dpds. In the Atharvan we have dpd'm, vi.119.3: 
xix.4.2 (and so Ait. Br. iii.46). This is by origin perhaps a con- 
tract accusative for dpd(s)am ; it is, at any rate, the starting- 
point of the transition. As though the ace. were dpd'-m, a 
N.s.f. dpd' is formed (Cat. Br. ii.1.3 4 ), and so all the later forms 
from an d-stem. C£ as-stems, A.s. 

Prom nds the sanhitas have nas-d', -J, -6s ; and to this same 
stem I refer nd's-d, RV. ii.39.6, as a regular strong form (see 
p. 361 ; Bollensen, Orient und Occ. ii.474, reads nd'se). I rec- 
ognize nd'se of the AV., v.23.3, as the first transition-form, an 
A.d.f. of stem nd'sd. Just such a strong form as nd's-d mediates 
the transition from stem nd's to stem ndsd. 

From stem kd's, f., we have kd's-am, kds-d', -as. A transition- 
vocative Jcdse (stem kd'sd) is seen AV. vi. 105. 1,2,3. 

As a transition-stem, from nimish, we must assume nimisha 
(distinguished from the organic formation nimeshd by its accent 
and lack of guna). This nimisha we have, with regular 
bahuwihi accent, in a-nimishd-s (x.103.1), -dm (i.24.6), -i,na 
(x.103.2), -d's, (ii.27.9). Cf. Garbe, Kuhn's Zeitsch., xxiii.492. 
In d-nimish-am and -d we have regular sA-forms. 

From ish-ds comes ishd-m (21) ; from ds-d', *dsdyd, adv. dsayd' 

( 2 )- 
Transition to the ^-declension. The form ishdye (accent — cf. 

ish-i) is a transition-dative to ish. See BR. i.826-7, and cf. Del- 

bruck, Verbum, § 207. On the other hand, the stem tv'ishi is 

probably an organic formation independent of tvish. 

The stems mdnsd and md'ns, n., are supplementary; but in 
this case the vowel-stem is probably the older (cf. vandhurd, 
vandhiir, p. 486) ; otherwise, its oxytone accent is problematic. 

Irregular accent. The following accusatives p. are oxytone : 
(masc.) jndsds, punsds, mdsds; (fem.) ishds, ushds, dvishds. The 
A.s.f. kds-dm rests on an error of the MSS. Gr. is wrong in 
taking md'sas as G.s. ; it is a N.s. transition-form, md'sa-s. 
Adverbial displacement is seen in d- and upa-vyushdm and 
dprdvrshdm. 

The diaskeuasts take account of the metrical value of y as a 
syllable in the accent of the vocatives dyails, jyd"ke, etc. (see 
p. 432) ; and so also in vy&sh-i, pronounced vitish-i. If the stem 
were a true monosyllable, it should be written vytishA. Cf. prd'p-i 
(==.pra-\-d$-i). In presh-d', on the other hand, the dissyllabic char- 
acter of the stem (pra-\-ish-d) is disregarded ; and so, naturally 
enough, in bhaas-d'. 

Nominative Singular Masculine and Feminine. 

A. Masculines: edhamdna-dvU (indras); an-dh (stem an-dksh). 
Cf. dadhrk (stem dadhrsh)=dhrshtas, BR. v. 1485. 

Feminines: vipr&t, AV. xx.134.4; prdvrt, Kathas. ii.56. The 
form idd serves as N.s. to ish. 



N.s.m.f.] Noun- Inflection in the Veda. 495 

B. Masculines: sudd's; ddre-bhds, i.65.10. Gr. refers this to a 
stem ddrb-bhds. Of course, it cannot be a compound of a verbal 
-bhd, because it is barytone ; but if taken as bahuvrthi compound 
of a substantive bhd, the accent would be correct, and dtire-bhds 
a correct N.s.m. of ddre-bhd. See BR. v.234. Here Gr. puts 
ayd'[s], i. 87.4 b, ayd' tpdndh, p. ayd' ' ; see p. 358 end. Pumdn, 
for pumdhs, occurs with elided * (cf. p. 346 med.), before iti, v. 61. 8; 
before e-, x.130.2; before a-, iv.3.10; pumdnp ca, AV. x.4.8 ; 
pu-rndn, RV. x.130.2 and vi.75.14 ; ptimdn, AV, 6 times before 
vowels (as, i.8.1 : iii.23.3) and 5 times before consonants. The 
iiom. (sa drksho) mds is found ; see Weber, Naxatra ii.281. 
For ghrdns (BR. stem ghrdns), see aw-stems, N.s.m. 

Feminities: djnds, x.39.6 ; dpt's, x. 128.3: AV. ii. 29.3. 

Transitions to the a-declension. Masc. : md'sa-s, x.85.5d, 'Of 
years the month is the constituent part ;' animishd-s. Fem. : 
dpd'. See p. 494. 

Nominative and Accusative Singular Neuter. 

Here belong: bhd's, 6 (and AV. vii.14.2); bhdas or bhdds, 2; 
do's ; yds, 20; pamyds, i.34.6; md's, 'flesh,' iv.33.4: v. 29.8 ; 
dadhr'k, as adv., 3 (see BR. v. 1485). 

Accusative Singular Masculine and Feminine. 

A. Masculines : ghrtaprusham, 2 ; jaradvisham ; brahma- 
dv'isham; vipodpusham. Feminities: dnimisham, 2; isham, 75; 
dvisham; pr'ksham; from AV., 'isham, 9 ; doshani- and hrdaya- 
prisham, vi.9.2. 

Adverbial shift of accent is seen in: dvyushdm, AV. iv.5.7; 
dprdvrshdm, Cat. Br. v.5.2 3 ; upavyushdm, TBr. i.1.9"; 5.2'. 

B. Masculines : p&mdnsam, 3 (and AV., 8) ; ayd'sam, 2 ; uktha- 
pd'sam (p. d) ; md'sarn; sudd' sarin, 7 ; subhd'sam; sudpisham; 
sud'sam. 

Feminines: dpisham, 2; prapisham, 2 (and AV. xiii.4.27) ; 
kd's-am, AV. v.22.11 (MSS. and ed. kdsdm). 

Transition-forms : animishdm; dpd'm; see p. 494. 

Instrumental Singular Masculine, Feminine, and Neuter. 

A. Masculine : vipvdp-dshd. Feminines : d-nimishd, 2 ; ishd' 
23; tvishd', 2; from A V., ishd', 4; prdvr'shd, xix.6.11. Neu- 
ters: ghrtaprdshd; presh-d', ix.97.1 (should be preshd; see p. 494). 

Transition-form: a-nimishena; see page 494. 

B. Masculine: mdsd', 2. Feminines: abhipdsd; dpdsd, 4 
dplshd; nihpdsd; pds-d', x.20.2 : vii.48.3; from AV., avapdsd 
nihpdsd, pardpdsd, vi.45.2 ; dpdsd, vii.57.1 ; dpishd, xiv.2.9 
praplshd,vi. 133.1 : xiv.1.53; nasd', ii.27.2 : v. 14.1; kdsd', v.22.10 
Grassmann's article pdsd', f., Wb. 1393, may be struck out . 
pdsd'm, ii.23.12, may be taken as G.p.m. of pdm (see p. 353); and 
pdsd', vii.48.3 (Gr. pdsd'[s\— see p. 363), is an I.s.f., as in x.20.2. 

Neuters: dsd', 22 ; bhdsd', 6; bhaasd', 3. 



496 C. R. Lanman, [Bad. sh, a-stems. 

Dative Singular Masculine and Feminine. 

A. Masculines : rshidvishe; brahmadvishe, 5 ; gavishe; papu- 
ishe; sdkam-tikshe; dvidvishe, AV. i.34.5. Here belongs perhaps 
aram-ishe, RV. viii.46.17, text dram ishe. 

Feminines : tvisM; prkshe", 2 ; rishk, 2 ; ishe, 15 (and AV. 
vii.78.1). 
Transition to the i-declension : ishdye, vi.52.15. 

B. Masculines: sudd'se, 16; dprdyushe, i.127.5; punse, AV. 
v.30.3. 

Ablative Singular Masculine, Feminine, and Neuter. 

A. Feminines: abhiprishas; tvishds, 2; dvishds, 3; nimishas; 
rishds, 12; ishds, ix.79.1. 

B. Masculine: punsds, vr.61.6 : AV. iii.6.1: xii.3.1. 
Feminine : kdsds, AV. i.12.3. Neuter : dsds. 

Genitive Singular Masculine and Feminine. 

A. Masculine : prAgavrshas (napdt). Feminines : abhra- 
prdshas; ishds, 10; ishas {pate); ushds, i.69.1,9: vii.10.1 (and 
AV. xvi.6.6fa'a); ushds-ushas ; dvishds, 3 times; prkshds, 4; 
nimishas, ii.28.6. 

B. Masculines : apdsas; sudd'sas, 5 ; sud'sas; punsds, 2 (and 
AV. vi.70.1-3). The accent forbids our taking rnd'sas as G.s. ; 
see p. 493, and N.s.m. 

Locative Singular Masculine and Feminine. 

A. Feminines : nimishi, 2 ; viiishi, 5 ; prdvr'shi, 2 (and AV. 
xii.1.46). 

B. Masculines : mdsi-mdsi; mdsi, 2 ; from AV., mdsi, 7 ; punsi, 
vi.11.2. Feminines: updsi, 2 ; from AV, dpishi, v.24.1-17 ; 
prapishi, v. 12.11 ; nasi, VS. xix. 90: xxi.49. 

Vocative Singular Masculine and Feminine. 

B. Masculine : pumas. Feminine : kdse, transition-form 

(p. 494) ; similarly, ide (p. 493). 

Nominative and Accusative Dual Masculine and Feminine. 

B. Masculines : ukthapd'sd (p. a) ; sil'rydmd'sd, 5 (can also be 
referred to stem -ind'sa). Feminine : nd's-d, ii.39.6 (p. 494). 
Transition-form : nd'se, AV. v.23.3 ; see p. 494. 

Genitive and Locative Dual Feminine. 
B. Genitive: nasds. Locative: nasds, AV. xix.60.1 : TS. v.5.9'. 

Nominative and Vocative Plural Masculine and Feminine. 

A. Masculines : asaca-dvishas; anrta-dvishas, brahrna-; ghrta- 
prushas, 2 ; pari- ; gavishas, 2 ; brhadukshas ; mdkshas, 2 ; 
mU'shas, 2 ; vd'tatvishas; yajna-mii'shas, TS. iii.5 Aquater. 



N.V.p.m.l] Noun-Inflection in (he Veda. 497 

Feminines : ishas, 9; ghrtaprushas ; tvishas; nemann4shas ; 
patidvishas; pr'kshas, 7 ; samishas; sdkam'&kshas; supr'kshas; 
from A V., dnddhrshas, vi.21.3 ; nimishas, iv.16.5 ; vy&shas, 
xiii.3.21. 

B. Masculines : aydsas; ayd'sas, 8 ; ukthapd'sas, p. d, (see p. 
493), 3 ; susampdsas, AV. xviii.3.16. 

Transition-forms : md'sds; animishd's; see p. 493-4. 

Feminines : dpdsas, 5 ; prapishas; dpishas, 3 (and AV., 4) ; 
from AV., pra-pishas, vi-, sam-, xi.8.27 ; vi-srdsas, xix.34.3 ; su- 
vii.76.1fo's/ svayarh-, 2. 

Accusative Plural Masculine and Feminine. 

A. Masculine : brahmadvishas, 6. ITeminines : ishas, 63 ; 
camrishas; pr'kshas, 9 ; dvishas, 39 (and AV. ii.6.5) ; viprushas, 
AV. ix.5.19. ' 

Irregularly accented : ishas, 7 ; ushds, 2 ; dvishas, 4. 

B. Masculines : and'sas; drv&rdhvabhdsas; ayd'sas; apdsas; 
hrtsudsas. Feminines : dpdsas; dpishas, 2 (and AV., 2) ; 
prapishas. 

Irregularly accented: jndsds; mdsds, 5 (and AV. viii.9.17) ; 
punsds, 3. 

Transition-form : tnd'sdn, 2 ; see p. 493. 

Instrumental Plural Masculine and Feminine. 

A. Feminine : viprudbhis, VS. xxv.9. The form iddbhis (6), 
serves as an I. p. to ish: see pp. 493 and 483. 

B. Masculine : mddbhis. 

Dative and Ablative Plural Masculine and Feminine. 

A. Feminine: viprudbhyas (abl.), Qat. Br. ix.1.1 16 . 

B. Masculine: mddb'hyds (dat.), AV. iii.10.10; (abl.) xix.27.2. 

Genitive Plural Masculine and Feminine. 

A. Feminines : ishd'm, 2 ; ishddm, 6 (resolution unnecessary in 
viii.46.26) ; dvishd'm; viprushdm, Qat. Br. iv.2.5'. 

B. Masculines : ayd'sdm, 2 ; vasd'rn (or f.) ; mdsd'm, x.138.6 
(and AV. i.35.4) ; punsQlm, AV. iv.4.4 : vii.13.1 : xx.l 29.14 ?. 

Locative Plural Masculine, Feminine, and Neuter. 

Here belong : punsu, AV. xii.1.25 ; mds-su, Pancav. Br. iv.4.1 ; 
mdsu (! sic — cf. as-stems, L.p.), TS. vii.5.2*. The neuter dohshu 
occurs Bhag. P. i.15.16. The form iddsti serves as a L.p. to ish; 
see p. 493. 

STEMS IN RADICAL H. 

Here belong : the compounds of -dr&h, -vdh, and -sdh; the mas- 
culines, sdh, go- and prati-dHh, and puruspr'h; the feminines; 
unhf, guh, druh, ndhf, nth, mih; ruh, d-ruh, upd-, pard-, pra-; 
sam-dih; ghrta- and sabar-duh; akshd- and upd-ndh, parindh; 
manomuh; gartdr&h; dnapasprh, puruspr'h; visr&h; ushnih. 



498 0. R. Lanman, [Radical 

Compounds of -druh: a-, akshnayd-, anabhi-, antaka-, abhi-, asma-, pv/ru-, vishu-. 

Compounds of -vdh : anad-, indra-, girva-, dakshind-, prshti-, madhywma-, vajra-, 
vtra-, saJwr, sttshtu-, svasti-, havir-, ham/a-, hoiira-. 

Compounds of -sah : carshani-, dyumtid-, dhanvd-, pra-, prd-, yajnd-, raihd-, 
vibhvd-, vifvd-, satrd-, sadd-. Compounds of -shah, p. -sah : abhimdti-, abhi-, rtt-, 
rshi-, jand-, turd-, nih-, rtt-, nr-, purd-, prtand-, prdfu-, bhuri-, rayi-, vane-, vlrd-, 
vipvd-, vrthd-, (atr-Or, satrd-. 

Neuter forms occur from the stems puruspr'h and mdh. For 
stem dadr'h?, see BR. iii.505 : v. 1485. 

For sardh, see grA-stems, p. 454 ; for ndh, dh-stems, p. 478, and 
p. 483 ; for dnh, see Ab.s.f. The derivation of ugdahak (stem 
upd-dah?) is uncertain. 

The peculiar compound of anas and -vdh makes its strong forms 
from the stem anadvd'h; the weak, from anadHth; the middle, 
from anadtit; but see N.s.m. 

The distinction of strong and weak forms is seen further as 
follows : in the compounds of -vdh, strong -vd'h ( pada always 
-vd'h); and in the compounds of -sah, strong -sd'h (pada always 
-s&'h). This lengthening occurs in the N. and A.s.m. and N.p.m. ; 
and, besides, in the following forms : indravd'hd, anadvd'hau, 
indravd'hau, dakshindvd't (N.s.f.), and havyavd'ham (A.s.f.). 

One cannot help asking whether we have in -sd'h a metrical or a declensional 
lengthening. The long d seems sometimes to be due to the metre (e. g. in ii.4.9 : 
vi.7.3 ; 69.4, where d is penultimate of apdda of 11 syllables), and its suppression 
in strong cases is certainly brought about by the demands of the cadence ; but d 
also occurs when not favored by the metre (e. g. i.T9.86), or even when forbidden 
by it (e. g. viii.81.1a/). Indeed, there seems to be a good deal of arbitrariness 
about the entire matter (cf. v. 10.1c with vi.44.4c); but we must, on the whole, 
call it a declensional phenomenon. Compare Rik Pr. ix.15,26. 

The lengthening of the a of -sah is suppressed 18 times in strong 
cases. We can hardly call these " weak forms in place of strong," 
because the shortening may be referred to the metre in every 
instance. The & occurs : as penultimate of an anushtubh-pdda 
(i.8.1: v.23.2; 35.1: vi.14.4 : vii.94.'7 : viii.1.2 ; 26.20; 57.1; 
77.1 ; 87.10: ix.98.1) ; as penultimate of a jagatPpdda (i.119.10: 
vi.46.6 : viii.19.35 ; 21.10) ; as fifth of an anushtubh (i.l29.4c ; 
v.10.7) ; as ninth of a jagati (i.129.46). In all but the last three 
pddas the short is absolutely required by the cadence. 

On the other hand, the long d appears in weak cases twice : 
viz., hi i. 91. 185, where the demands of the penultimate excuse the 
long d (so vd'h-e, vii.24.5a) ; and in ii. 2 1.2(7, where there is no 
justification whatever for it. Cf. dvipd'de, p. 470. 

Transitions to the «-declension. I am inclined to regard mahd-s, 
-dsya, -e, -d'ni, -d', and -d'ndm as forms of transition from the weak 
cases of mdh. They belong exclusively to the Veda, and are 
used, especially mahd-s and mahd'ni, to supplement the defective 
mdh. From the weak stem anad-tih comes the transition -form 
dhenv-anaduhayos, Katy. Or. vii.2.23. 

Transition to the ^'-declension. The solitary inf. mahdye, x.65.3, 
is a transition-dative to mdh, equivalent to mah-e (as in viii.56.8). 
For accent, cf. mah-L 



A-stema] Noun-Inflection in the Veda. 499 

Irregularly accented are mahds, N.p.m. ; druhds and mahds, 
Ajp.m. For gHh&, see I.s.f. ; and for dnh-as, as-stems, Ab.s.n. 
Mah-f, du.n., is regularly accented, as a weak case. 

Transfer of aspiration occurs in a-dhruk, akshnayd-, antaka-, 
abhi-, asma-; go-dlvdk, prati-, sabar-; ucd-dhak(?). These, with 
ushar-blvut (p. 475), are all the instances of the kind in the declen- 
sional forms of the Rik. Somewhat akin to this is the nati of the 
s of -sdh even after an d when the aspiration disappears at the 
end of the word. Thus we have satrd-shd't and vipvd-shd't; but 
satrd-sd'ham, -sd'he, and vipvd-sd'ham. Irregular is prtand- 
shd'ham, vi.72.5. 

Nominative Singular Masculine and Feminine. 

I. Masculines : -shd't, with abhi-; rshi-; jand-; turd-, 4 ; nish-, 
2 ; purd-; prtand-, 5 ; prdcu-; bh&ri-; rayi-, 2 ; vane-; vlrd-; 
vrthd-; satrd-; from AV., abhi-, 2 ; turd-; ni- and patnt-, v.20.11 ; 
vipvd-, xii.1.54 : xiii.1.28 ; vrthd- f, xx.128.13. Here belong 
further : shd't, RV. i.63.3 ; rnadhyama-vd't; havir-vd't; havya- 
vd't, 9 times ; turya-vd't, TS. iv.3.3 s ; tury'a-vd't and ditya-vd't, 
iv.3.5'; 7.10 1 . Cf. rad. )'-stems, N.s. 

Feminines : dakshind-vd't; for sardt, see p. 454. 

Pashtha-vd't occurs TS.'iv.3.3 s , 5 1 '; 7.10 1 . In explanation of 
the dental t, Weber (Ind. Stud, xiii.107) suggests that the second 
part of the word was confounded with the suffix -vant. 

To precisely the same confusion he refers the N.s.m. anad-vd'n, 
AV. iv.ll.lquater, 2,4,10: v.17.18: viiL5.11: ix.1.22 : x'i.4.13 ; 
5.18:_xii.3.49: xix.39.4 : TS. iv.3.5 1 ; -vd'np ca, VS. xviii.27. 
Bohtlingk, however, suggests with much plausibility, the analogy 
of the perf. part, in -vdns, whose stems also have a triple form ; 
see Die Declination irn Sanskrit, § 70. 

II. Masculines : -dhruk, with a-, 3 ; akshnayd-; antaka-; abhi-; 
asma-, 3; upd-dhak?, 3; go-dh-uk ; prati-dhtik, AV. ix.4.4. 

Feminines: gartd-ruk; sabar-dhuk; ushnik, AV. xix.21.1. 
Transition to a-declension. Mahd-s (18) supplies a N.s.m. for 
mdh. 

Nominative and Accusative Singular Neuter. 

Here belongs puruspr'k. For dadhr'k, see «A-stems, A.s.n. 

Accusative Singular Masculine and Feminine. 

I. Masculines : -shd'ham, with abhimdti-, 2 ; rtt-; nr-; prtand-; 
-sd' ham, -with pro-; yajnd-; vicvd-, 3; satrd-,' 4; -v'd'ham, with 
vira-; suasti-; havya-, 20 (and AV, 3) ; hotra-; anad-, x.59.10 
and AV. ix.5.29 : xii.2.48. Feminine : havya-vd ham, x.8.6. 

The lengthening is suppressed : rti-shd' ham, 3 (cf. rtishd'ham) ; 
prtand-, 2 ; carshan%-sd' ham, 5 ; prd-, 2 ; vibhvd-, 2 ; sadd-. Cf. 
carshani-sdhd, -sdhas, rathds&'hd, and. p. 498. 

II. Masculines : adruham, 2 ; puruspr'ham, 14. 
Feminines : gtiham; druham; parindham; tniham, 6. 



500 0. R. Lawman, [Rad. ^-sterns. 

Instrumental Singular Masculine, Feminine, and Neuter. 

Masculines : dhanvdsdhd; puruspr'hd; vishudruhd. 

Feminines : druhd', 2 ; prdsdhd, 2 ; mahd', 2 (cf. mahyd') ; vi- 
srtihd; guh-dl , i.67.6. Gtihd, occurring 53 times, may be an 
adverb with recessive accent, from guh-dl , like divd from div-d'; 
or a homophonous I.s.f. (see p. 358). Neuter : mahd', 4. 

Dative Singular Masculine, Feminine, and Neuter. 

Masculines : adruhe; abhidri'ihe; carshanisdhe; druhe (or f.) ; 
mah'e, 42 (and AV. ii.5.4 : v.11.1); satrdsd'hef, p. ti, ii.21.2 (see 
p. 498). 

Feminines : goduhe; make, viii.26.2 ; 46.25 : ix.44.1. 

Neuter : mah'e, 33 ; AV. x.6.8 : xix.24.2,3 ; RV. vii.30.1<7, text 
mdhi. For vd'he, see Delbrttck, Verbum, p. 222. 

Transition-form : mahdye (inf.), x.65.3. 

Ablative Singular Masculine, Feminine, and Neuter. 

Masculines: rtishdhas ; druhds, 3; mahds, 5. Feminine: 
druhds, 5, and AV. vii.103.1. For dnh-as, vi.3.1, see as-stems, 
Ab.s.n. Neuter: mahds, 11, and AV. vii.26.8. 

Genitive Singular Masculine, Feminine, and Neuter. 

Masculines : adruhas; druhds, 3 ; prtandshdhas ; mahds, 61 ; 
abhimdtishd' has !, p. &, i.91.18 (see p. 498) ; from AV., anad&has, 
iv.ll. 8,9,11; mahds, iv. 1.6 : x.7.2. 

The transition-form mahdsya occurs i.150.1, q.v. 

Feminines: prdsdhas; mihds, 2 ; druhds, 3 (and AV., 11). 

Neuters : puruspr'has; mahds, 39 ; mahas (with voc.), vii.64.2. 

Locative Singular Masculine and Feminine. 

Masculine: anaduhi, AV. iv.ll. 1. Feminines: updndhi, 
xx.133.4; parindhi, xix.48.1. The transition-form mah'e occurs 
as m., ix.66.13 ; and as n., 5 times and in ix.110.2. 

Vocative Singular Masculine and Feminine. 

Masculines: havyavdt; from AV., prtandshdt, v.14.8; godhuk, 
vii.73.6 (ed. nom.). Feminine : ddhruk. 

Nom., Aoc, and Voc. Dual Masculine and Feminine. 

I. Masculines: indra-vd'hd (d), 2; carshanisdhd ! ; rathd- 
sd'hd !; adruhd, 2 ; dnabhidruhd; puruspr'hd. 

Feminines: adruhd; adruhd, 1. 

IL Masculines: indravd'hau; anadvd'hau, x.85. 10: AV. iii. 
11.5: vii.53.5. 

Nominative and Accusative Dual Neuter. 
Here belongs mah-t', il21.ll : ix.68.3. 



N.V.p.m.f.] Noun- Inflection in the Veda. 501 

Nominative and Vocative Plural Masculine and Feminine. 

I. Masculines : abhimdtishd' has, 3 ; patrtishd'has; -vd'has, with 
indra-j vajra-; vira-; saha-; sushthvr; havya-; anad-, AV. xix. 
50.2 ; prshti-, xviii.4.10. Gr. reads girva-vd'has, vi.24.6 (see p. 
353 med.). ' 

II. Masculines : carshanisdhas; adruhas, 3 ; purusprhas; puru- 
sprfhas, 2 ; god'dhas; dr&has; adr&has, 9, and AV. vi.7.1. 

Irregularly accented is mahds, iii.7.6 : x.61.27 ; 77.8. 

Feminines: adriihas, 2; ghrtadtihas; purudr&has; puruspr'has, 
3; rnihas, 2; r'&has; visrtihas; from AV., manom/uhas, ii.2.5; 
rhhas, d-rdhas, pra-, xiii.1.9. 

Nominative and Accusative Plural Neuter. 

There are no forms like -dunhi, -vdnhi. Their place is supplied 
by transition-forms. Thus instead of manhi, we have mahd'ni 
(11), mahd'. 

Accusative Plural Masculine and Feminine. 

Masculines: adr&has; puruspr'has; anaduhas, AV. iv.11.3. 

Irregular accent is seen in druhds (vii.104.7) and mahds (20 
times). Perhaps mahds is so accented to differentiate it from 
mdhas, 'might.' Cf. apds ('aquas') and dpas ('opus'). 

Feminines: akshdndhas ; adr&has; updrtihas; d/rdhas, 5; 
mihas; samdihas; from AV.,drtihas, xix. 15.2; r&has,xiii. 1.4,8,26; 
3.26; prardhas, xiii.1.8; nihas, ii.6.5 and VS. xxvii.6. 

Dative Plural Masculine and Feminine. 

Masculine : anadtidbhyas, AV. vi.59.1. Feminine : for 

sarddbhias, see p. 454 ; for nadbhids, pp. 478 and 483. 

Genitive and Locative Plural Masculine. 

Genitive : mahd'm, iv.5.9 : ix.109.7. Roth suggests that 
mahd'm in ii.24.11 may be G.p. of a noun mdh (root manh), 
'gift.' 'Und der du dich frohlich an den Gaben kraftig gestarket 
hast.' So iv.23.1. In x.46.55, the resolution mahddm is needless. 
In vi.66.3, Gr. reads mahddm, text mahd. The transition-form 
mahd'naam occurs five times. 

Locative: anadtitsu. 



STEMS WHOSE SUFFIXES END IN T, N, A.ND & 
According to the plan proposed on p. 327, we now come to 
the stems formed by derivative suffixes ending in t, n, and s. 

STEMS IN VOCALIC ROOT +T. 
The stems ending in the quasi-suffix t are declined just like 
those in radical t. It may fairly be questioned whether the 
added t ought not rather to be called a root-determinative. 
vol. x. 67 



502 C. R. Lawman, [Stems in 

This t is added to the roots kr (properly kar, etc.), kshi, gu, ci, 
cyu, ji, dhr, dhru, dhvr, pi, pru, bhr, mi, yu, ri, vr, pri, pru, su, 
sr, stu, spr, sru, hu, hr, hru, hvr; that is, to roots in i, H, and r 
(ar), so that almost all compounds of roots in these short vowels are 
thus thrown into the consonant declension (pp. 453, 367, 401). 
It is properly added only to roots in these vowels ; but we have 
the tertiary formation (adhva-gd ', -gd') adhva-gd'-t, and similarly 
dyu-gdt, nava-gdt, and (sam-hdn, -hd') samhd't. Cf. Benfey, 
Oram. p. 131, 1. 26. Neuters are very rare (stems upashpit, tawtt- 
kr't, dirghaprM, dyugdt, navagdt, vipvajit). 

For the accent of parihvrtd ', see I.s.f. ; for hrut-ds, A.p.m. 

I do not think that there are any instances of transition to the a-declension on 
the part of the stems here treated. Their last element is almost invariably an 
active verbal, and if they went into the a-declension, they would coincide in form 
with passive participles and their compounds, but would conflict with them in 
meaning. 

Nominative Singular Masculine and Feminine. 

Here belong 79 masc. forms (from 50 stems) ; and 4 fem. forms 
(from 3 stems). 

Masculines: -kr't, with amkdma-; abhishti-, 3; dji-; icdna-, 6; rshi-, 2; Tchaya-, 
2; jyotish-; tani- ; paihi-, 4; puru- ; bhadra-; uloka-,2; vayas-,2; varivas- ; 
vijesha-; su-, 2; steya-; siona-; — -jit, with apsu-, 3 ; at-; acva-; go-, 2; dhana-; 
ranya-; raiha-; vipoa-, 2 ; samsrshta-; satrd-, 2 ; samarya-; sahasra-, 2 ; suar-, 2 ; 
hiranya-; — acyutacyut; dks'hit) rnatit; rtacit, 3 ; Mbishaspr't ; carshantdhr't, 
3; dirghacrut; dhanaspr't; parihrut; bhdrabhr't; madacytit; madhusMf, 3; 
yakshabhft; vajrdbhr't, 2; vdjasr't; somas&t; svasr't; havanacrut; from AV., 
dddrasrt (i.20.1), etc. 

Feminines: upamit, 2; rtacit (iv.16.10); pracastikr't. 

Nominative and Accusative Singular Neuter. 

Here belong: vipvajit; dirghaprut (pdrma), vii.16.8 ; navagdt, 
TS. iv.8.11 1 : AV. iii.10.4; — and, used as adverbs, upashttit; 
dyugdt; sakr't, 7 ; dirghaprut, vii.61.2 (so BR. ; see A.p.n.). 

Accusative Singular Masculine and Feminine. 
Here belong 46 masc. forms (from 25 stems) ; and 5 fem. forms 
(from 5 stems). 

Masculines : adhikshitam ; arnovr'tam ; ukthabhr'tam ; vdaprutam, 2 ; rnacyut- 
am ; carshanidhr'tam, 2 ; tivrasiOam ; trshucyutam ; dirghacrttam ; dushkr'tam ; 
dveshoyutam, ; dhanaspr'tam, 8; nadivftam, 2 ; parisiidam; prdtarjitam; mada- 
cy&tam, 8 ; mitMkr'twm ; gravojitam ; sarhjitam, ; satyadhvr'tam ; sdmabhr'tam ; 
sukr'tam, 3 ; svarjiiam ; havanagrvtam, 2 ; havishkftam, 2.^ 

Feminines: aramkftam ; vpaprutam; devapr&tam ; niyutam ; parisr&tam. 

Instrumental Singular Masculine and Feminine. 

Masculines : carshanidhr'td; brahmakr'td, 2 ; somas-utd. 

Feminine : niy&td, 2. The accent of parihvrt-d', viii.47.6, is 
quite anomalous ; cf. vimrdhd, p. 477. It may be L.s. of pari- 
hvr-ti : ' For in the very act of deceiving you, a man loses the 
blessing ye already have given.' Stems of this form regularly 
accent the preposition (pdrihvrti) ; but cf. d-sakt'i and d-suti. 



voc. root+£] Noun- Inflection in the Veda. 503 

Dative Singular Masculine. 

Here belong : -kr'te, with aram-, 2 ; dush-, 2 ; dharma-; brahma-; 
an-, 13 ; — -jite, with a b-; apva-; urvard-; go-; dhana-; nr-; 
vipva-; satrd-; mar-; — girikshite; parvatacytite. 

Ablative Singular Masculine and Feminine. 

Masculines : brahmakr'tas; sarvah&tas, 2. Feminine : abhi- 
hr&tas, 2. 

Genitive Singular, Masculine and Feminine. 

Masculines : carshanidhr'tas, 2 ; samjitas; aukr'tas, 3 ; havish- 
kr'tas, 2. Feminines : ishastutas; devastMas; parisrutas, AV. 
iii.12.7. 

Locative Singular Masculine. 

Here belong : rnadacyuti ; go-, vasu-, sarhdhand-jiti (AV. 
xiii.1.37). 

Vocative Singular Masculine and Feminine. 

Masculines : acyutacyut; urukrt; rtacit; khayakrt; purukrt, 
5 ; ranakrt; vandanaprut; sahasrajit, 4. Feminines : from 

AV., upabhrt, xviii.4.6 ; rd'shtrabhrt, vi.118.2. 

Nom., Aoc, and Voc. Dual Masculine and Feminine. 

I. Masculines : gofitd; diviksMtd; pariksMtd, 2 ; madacyutd, 
3 ; as voc, 2 ; supr&td; havana<yr&td; as voc, 3 ; pdthikrtd, AV. 
xviii.2.53. For ishukr'td, see pp. 466, 468. 

II. Masculines : ddushkrtau; krshnaprutau; sakshitau; from 
AV., sa-kr'tau, xi.1.10 ; su-, xii.3.44; bhadra- and su-pru-tau, 
xvi.2.4. 

Feminine: ugrqjitau, AV. vi.118.1. 

Genitive Dual Feminine. 
Here belong : sukr'tos; parikshitos. 

Nominative and Vocative Plural Masculine and Feminine. 

Here belong 49 masc. forms (from 32 stems) ; and 24 fem. forms 
(from 7 stems). 

Masculines : -kr'tas, with aram-, 2 ; fydna-; dudhra-; dush-; dhdnid-; brahma-, 
3 ; yajria-nish-; vayas-, 2 ; su-, 8 ; havish-; — -cyuias, with dhantta-; dhruva-; 
parvata-; mada-, 2; — -kshitas, with apsu-; upa-; sa-; — udapr&tas, 3 ; 




7.23-2?), etc. For devastiias, see G-.s.f. 

Feminines: udaprutas; niy&tas, 14; parisrutas; pitubhr'tas; mitas ; sasrutas, 
2 ; stutas, 4. 

Nominative and Accusative Plural Neuter. 

See page 474. Here we may put dfrghaprut, viii.25.17 (vratd') ; 
vii.61.2 (mdnmdni). 



504 G. R. Lanman, [Participles 

Accusative Plural Masculine and Feminine. 

Masculines : apitas; gqjitas ; jyotishkr'tas / dushkr'tas, 3 ; 
nigdtas / madaey&tas ; sukftae, 3 ; hr&tas ; with irregular 
accent, hrutds, AV. vi.51.1. 

Feminines : Upastutas; dirghaprdtas; niytitas, 9 ; balihr'tas, 2 ; 
ritas; varunadhriitas; visr'tas; sarhhdtas (cf. p. 502) ; sasr&tas. 

Instrumental Plural Masculine and Feminine. 

Masculine : somastidbhis, 2. 

Feminines : antarikshapr&dbhis; niyAdbhis, 10. 

Dative and Ablative Plural Masculine and Neuter. 

Datives masc. : pathikr' dbhyas (Gr. -bhias) ; bandhurkshidbhyas. 
Ablative neut. : tanHkr' dbhyas. 

Genitive and Locative Plural Masculine and Feminine. 

Genitives masc. : mantrakr'tdm ; sukr'tdm, 2 ; from AV., 
ishubhr'tdm, iv.28.2 ; agnihotrahMdm, iii.28.6. 

Genitives fern. : abhihrutdm; niytitdm; from AV., rathajitdm, 
vi.130.1 ; upar, pari-, prati-niUdm, ix.3.1. 

Locatives: m., sukrtsu; f., paripritsu, Cat. Br. ix.1.1*. 

PARTICIPIAL STEMS IN NT AND T. 

A. The suffix ant or nt forms participles of the present, future, 
desiderative, causative, and denominative. The nasal appears 
only in the strong cases. In the weak it is regularly dropped, 
and if the last syllable of the stem had the accent, this is shifted 
to the case-ending. The category under which these participles 
fall I call for convenience " A." Many forms which, if simple, 
would be oxytone become barytone when compounded with a-, 
an-. Sthd, although a reduplicating root, has gone over to the 
a-conjugation, and we have the stem tishthant, with nasal. 

B. There is, however, a very considerable number of participial 
stems, which do not show the nasal even in the strong cases. 
These are the participles from reduplicated verb-stems, excluding 
the desideratives. They are regularly accented on the syllable 
of reduplication ; e. g. mb-mi-at, bd-ps-at. Accordingly, they can 
not shift the accent to the ending in the weak cases. These I 
regard as making up category " B." 

Besides these there are a few stems embraced in this section 
which are not properly participles. Brhdnt, rhdnt, pr'sliant, 
r(i$ant, and jdgat have exchanged their participial function for 
one almost purely adjectival or substantival ; but they retain the 
participial inflection. Mahdnt (s-d-mahant) has lost its participial 
meaning, and deviates markedly from the declension of participles 
in the strong cases, where it shows a long vowel. For ddnt and 
its compounds, see below. The four stems iyant and kiyant, 
i'vant and ki'vant, although in no wise participles, are yet de- 
clined like them, and so are treated here. 



in nt, <.] Noun- Inflection in the Veda. 505 

There are several participles which I have included in category 
B, because they appear in the strong cases without a nasal. They 
are: dd'patas, N.p.m., 2; pd'sat, N.s.m., 4; -atd, N.d.m., 2; -atcu, 
G.s.m. ; vdghdt, N.s.m. ; -dte, 3 ; -dtas, G.s.m., 5 ; -dtas, N.p.m., 5 ; 
-atas, V.p.m. ; -ddbhis, 4 ; -dtdm, 3. Here also is to be placed 
dd'sat: it occurs only in G.s.m., dd'satas; but the N.p.m. ought 
to be dd'satas. Finally, the aorist participle of dah, dhdkshat, 
omits the nasal: dhdkshat, N.s.m., 3; -atas, G.s.m.; and similarly 
ubhayd'-dat, N.s. ; -datas, N.p. 

Vdghdt is doubly peculiar : the nasal is lacking in the strong 
cases ; and the accent, although resting upon the final syllable of 
the stem, is not shifted to the ending in the weak eases. Grass- 
mann regards ddp, dds, and pds as reduplicated forms of ddp 
(vii.3.7, dd'pema), dds, and pds (pans), and interprets the absence 
of the nasal in the strong cases as evidence in support of his view. 
Possibly vdgh of vdgh-dt is to be referred in like manner to 
vd(g)h; but see Roth in Kuhn's Zeitsch. xix.220. 

On the other hand, one unquestionably reduplicated form, 
vavrdhdntas, admits the nasal contrary to rule. Even this is per- 
haps due to the metre and the formal parallelism of the verse 
(iv.2.17, q. v.). There are two other forms, cakrdn and cdkdn, 
in regard to which there is much uncertainty. 

The N.s.m. ca-kr-dn (x.95.12,13— referred by BR. v.1253, to cakrdnt) is in both 
verses followed by »-, and, if we allow a mistake on the part of the padakdra, 
may stand for cakrdt. The oxytone accent is still anomalous : might not cakrdt 
(cf. for accent rdrdn, i.122.12) be a conditional subjunctive, 'if he remember'? 

In x.29.1 occurs cdkdn. Delbriick, Verium § 225, takes it as N.s.m. of the 
participle of kan, and regards its n as radical. In this case it is a shortened form 
for cdkanat, and is comparable only with kdnikrat (for kdnikradat, from krand) ; 
but the accent is wrong. Gr. takes it as N.s.m. of the participle of kd, with 
stomas, in which case both accent and nasal are anomalous. Both anomalies are 
avoided by taking it with Both as 2 or 3 p. sing. aor. of kan : ' Wie (der) welcher 
in der Kufe (ist), so ist er niedergesetzt;' i. e. So wie der Soma, so ist der Lobge- 
sang hier aufgetragen ; ' er moge Gefallen erregen' (cdkdn). See ErU. zum 
Nvrukta, p. 94. The passage seems to have been a crux, since the time of (jakalya. 

These participial stems are declined only in the masculine and 
neuter, both genders being alike, save, of course, the N.A.n. The 
feminines are formed by adding i to the weak stem ; see p. 367. 
In category A, the masculines are very numerous, the neuters 
very few ; in B, the masculines are considerable, the neuters only 
about half a dozen. 

The etymology of ddnt, ' tooth,' is uncertain (from ad, da= 
day, danpf). It seems to have been declined originally as a 
participle. Like participles and unlike nouns, it shifts the accent 
to the ending in the A. p. dat-ds; but it is unlike participles and 
like nouns, in shifting the accent to the ending in dadbhis, dad- 
bhyds. Cf. a.ddt-e, like nouns. The forms are given under the 
different cases; dan (x.115.2), ddnt-am (iv.6.8), etc. 

In the later texts this stem shows forms of transition to the 
a-declension ; for this the A.s. ddnt-am serves as point of depart- 
ure. The Rik has only one transition-form, ddnta-s (vi.75.1I — a 
late hymn). The AV. has : ddntds, ddntais. 



506 C. R. Lawman, [Ptcp. in nt, t 

Besides vdghdt, there are only two instances — quite sporadic — 
in which an oxytone stem does not shift the accent to the ending ; 
they are acoddte (in the miserable hymn v.44), and rathirdydtdm, 
ix.93.4. For srav-dt-as and sapc-dt-as, see pp. 466, 467 ; does not 
vdgh-dt-as perhaps belong in the same category with them ? 

Nominative Singular Masculine. 

A. Here belong 753 forms (from 228 stems). Examples are : 
drean, 19 ; sdn, 38 ; ydn, 32 ; iehdn, 15 ; krnvdn, 24 ; jandyan, 
27; jdndn, 24; si! dan, 17; pd$yan, 19; ghndn, 17. Further- 
more occur : mahd'n, 77; sti-mahdn; a-ddn; ddn, x.115.2 ; patd- 
dan, AV. xiv.2.68 ; h'tranyadan, Ait. Br. iii.6 ; from RV. again, 
pticidan,v.'7.1: vii.4.2 ; $r'enidan, x.20.3. 

When the form in -an occurs before o or j- in the Rik, the usual method of 
sandhi is -am, c- (e. g. iii.54.2 : vii.87.1), -an j- (e. g. i.173.11 : x.45.6). In regard 
to the " inserted t " before s- (e. g. minvdnt sddma pwra eti, x.20.5), see Rik Pr. 
iv.6 ; Aufrecht, Rigveda?, preface, p. vi ; and Whitney, to Ath. Pr. ii.9. Instances 
of the collision of -an with initial t are comparatively rare : the method of sandhi 
in the Rik is -an t- (e. g. i.l 11.22 : ji.3.2) ; in the Atharvan, there are six instances 
of -ans t- (vii.13.1 ; 18.2 : xiii.1.32,39 : xvii.10 : xviii.2.20) to three of -an t- 
(xiii.2.33: xiv.1.5: xix.28.4); in the TS., there are but two cases of -ans t- (iv. 
7.13 6 : v.4.12 1 ); the sandhi is generally -an t-. See Whitney, note to Ath. Pr. 
ii.26: and note to TPr. vi.14. If the s is here the historical relic of the case- 
ending, we ought to find it oftenest in the Rik ; but the Rik has but one example 
of it, dvddans tvdm (Prat, iv.33), and that in the last verse of the second book, a 
very late passage. Is it not rather (unlike the s of td'iis) the product of phonetic 
dissimilation (-ant t- to -ans t-) ? so Bohtlingk, Die Declinatim, % 22, Anm. 2, and 
note 38, q.v. For the forms in -amn, cf. Ourtius, Studien, ii.165. 

In 46 instances mahd'n appears as mah&'ft: 31 times before a- (e. g. viii.90.11,12 : 
so sumahdH, vii.8.2) ; 6 before i- (e. g. i.8.5) ; 6 before u- (e. g. iii.36.5) ; in iii.59.5 
and iv.8.2 before d-; in iii.53.9 before r-. We find mahd'nc before c- in iii.55.9 : 
vi.38.3: viii.33.8: x.4.2 (Prat, iv.32). The resolution mahaan is inadmissible: 
in vi.25.1(Z, vii.52.3c, and x.46.1a, the sixth syllable is syncopated. 

Transition-form: ddnta-s, as N.s. to ddnt, vi. 75.11. 

In vii.4.2 and x.20.3, the metre is out of order. It is rectified, 
if we substitute the longer transition-form in -dantas for the 
equivalent shorter grammatical form in -dan; thus, sdm yd vdnd 
yuvdte p&cidantah (but cf. p. 348 top) ; and ydm [id] dsd' krpd- 
nidam \ bhdsd'keium vardhdyanti \ bhrd'jate [sd] $r'enidantah 
(trochaic — cf. Ueb. ii.514). 

Quite irregular are: ubhayd'-dat, TS. ii.2.6*Jte; -dam, AV. 
v.19.2. Benfey takes rnahd'm as N.s.m. in ii.24.11 : iv.23.1 : ix. 
109.7 ; see Goitinger JSTachriehten, 1878, p. 190. 

B. Here belong 178 forms (from 48 stems). 

Enumeration: dbibhyal; cdrkrshat, 2; cd'kacat, 9 ; karikrat, 2 (cf. dcarikrat, 
AT. xi.5.6); kdnikradat, 32 ; cikyat; cekitat; tdvitvat; dardrat; dadat,5; dedicat; 
di'diat,6; di'diat; ddvidyutat, 3 ; dadhat, 18; dMhuvat, 3 ; davidhvat, 2; nd'nad- 
at, 4 ; nannamat ; pdniphanat ; ba-ps-at (bhas),^ 3 ; bibhyat, 2 ; jdrbhwrat, 2 ; 
bibhrat, 11; bhdribhrat; memiat ; mdrmrjat; mdrmrcat, 2; rd'rajat; rerihat, 3; 
rdruvat, 11; roruvat, 2; vd'vadat; vevidat, 2; vivishat, 4; vdrivrjat, 2 ; co^ucat, 
10- cdnifcadat; seshidhat; sdnishyadat ; jdnghanat, 2 ; gMnighnat ; jahai ; juhvat, 
2;'fTomAV.,jd'grat(vi.ll5.2: xvi.7.10). etc. ; from the Rik, further, kdnikrat, 
ix.63.20; dhdkshat, vi.3.4; ddkshan nd, p. dhdkshat nd, i.130.8; ddkshat, p. dh-, 
ii.4.1 ; vdghdt, ix.103.5 ; cd'sat, 4 times. 



N.s.m.] Noun- Inflection in the Veda. 507 

Gr. considers kdnikrat as a shortened form of the usual kdni- 
kradat; of. Jcdnikran-ti, 3 p. sing. 

Here belong, according to Gr., cdkdn and cakrdn, which, 
although reduplicated, yet have the nasal. See p. 505. 

Nominative and Accusative Singular Neuter. 

A. Here belong 272 forms (from 31 stems). 

Enumeration : anat ; viandt ; aminat ; avianat ; sat, 30 ; dsat, 4 ; d'sat, p. &'-, 
2; ydt,2; ishndt ; i'shat (x.89.14); $at,5; eshatf; kuldydyat; carat,3; citdyat; 
Irpdt, 5; dhrshdt, 10; pdtat; patdyat; irfhat, 148; rnahdt, 26; rnindt; raghuydt; 
r&fat, 14; vdrdhat; cdnsat; fikshat; cvdyat; sdkshat; sunvdt; from AT., bhav- 
ishydt (6 times, as x.7.9), etc. ; from the Rik, further, iyat; i'vat, 2 ; Uyat, 2. 
For dyugdt, see p. 502. 

As adverbs, with adverbial shift of accent, occur : dravdt (drdv- 
ant), 5 ; drahydt (! dr'hyani), once. 

B. Here belong: ddvidyutat, x.43.4 ; pepipat, x. 12.7. 7; jdgat, 
16 times; ydyuvat, AV. xix.47.2. 

Accusative Singular Masculine. 

A. Here belong 315 forms (from 121 stems). Examples are : 
sdntam, 28 ; ydntam, 11 ; cdrantam, 24 ; brhdntam, 35 ; updntam, 
8. Furthermore occur : mahd'ntam, 8 ; ddnt-am (Gr. ddnta-m), 
iv.6.8 ; hiranyadant-am, v.2.3. For mahd'm, see osw-stems, A.s.m. 

B. Here belong: d-sapcatam (i.112.9 — seep. 467); gdnigmatam; 
dddatam; di' diatom; nd'nadatarn; pdnipnatam, 4 ; bibhratam, 
5 ; from AV., kdrikratam, x.4.13 ; bibhratam, xiv.2.30: xix.32.9 ; 
rdruvatam, xi.10.26. 

Instrumental Singular Masculine and Neuter. 

A. Masculine. Oxytones : satd'; jdnatd' ; tujatd', 2 ; brhatd', 
11; bhindatd' ; mahatd', 5; vanushyatd' ; datd! , x.115.2: AV. 
vi.56.3. Barytones : dghnatd; dvatd; dsunvatd, 3 ; cdratd; 
P'ushyatd; r&patd, 2 ; from AV, pydvddatd (vii.65.3), etc. 

Neuter. Oxytones : gavyatd', 3 ; tudyatdl ; dhrshatd', as adv., 
18 (and perhaps Val. 4.3) ; brhatd', 7 ; mahatd', 3 ; upatd', 2 ; 
pucatd' ; pravasyatd', 2 ; samaryatd' ; sumndyatd'. Barytones : 
ddrpyatd; d'satd, p. &'-; dsredhatd; dhedatd, 3 ; rupatd. 

B. Masculines: dddatd; pdpucatd. Neuter: jdgatdA. 164.25 : 
AV. iii.12.7. 

Dative Singular Masculine and Neuter. 

A. Masculine. There are 187 oxytone forms (from 38 stems), 
and 39 barytone forms (from 26 stems). 

Oxytones: avasyaU; avishyate; afvdyaU; sate, 5; yaU, 9; ishayate; ishudh- 
yate ; ish&yate ; rjuyate, 3; rtdyate, 6; rhate; hrnvaU, 2; gavyate; grqate, 29; 
jwate; jdnate; tudyatt; devayate, 5; dvishaU, 2;' dhiydyaU; dhAnvate; prnate, 
3 ; prchate ; brhati, 16 ; brahmanyate ; mahayatt ; mahate, 4 ; mdnavasyate ; 
yajndyaU; vanvate; Ufate, 6; vidhate, 15; gravasyaU, 3; friycaU, 3; sakhiyate, 
3; silicate; sunvate, 33; stuvatS, 21. 



508 C. R. Lanman, [Ptcp. in n(, t. 

Barytones: aghnate; acoddte (sic, v.44.2) ; dminate; drcate, 3; drhate; iyaksh- 
ate ; cdrate ; cetate ; idrate ; tr'shyate, 2 ; ddpate ; pdcate ; pipishate ; pushyate ; 
ydjate ; yudhyate ; rd'jate ; ri'shate, 2 ; v&date, 2 ; (annate, 3 ; fdrdhate, 4 ; sishd- 
sate; hdryate, 4; and, t'vate; kiyate ; addte, i.189.5. 

The utterly incorrect and impossible accent acoddte is particu- 
larly appropriate for hymn v. 44 ; d-codate would be correct. If 
we grant a stem acoddnt, the D.s. would be -atk, never -dte. 

Neuter. Oxytones : brhate; mahat'e, 10 ; pucat&; sandyate. 
Barytone: i'vate. 

B. Masculines : jtijoshate; dddhate; bibhrate; sdpcate; ghdni- 
ghnate; jtihvate; vdghdte (accent 1 see p. 505), 3. Neuter : 
jdgate, AV. i.31.4. 

Ablative Singular Masculine and Neuter. 

A. Masculines: krnvatds; prnatds; brhatds, 13 ; vanushyatds, 
4; and, dbhufijatas; dvadatas; ri'shatas,S; jighdnsatas. Neu- 
ters : brhatds, 3 ; mahatds; and, dsatas; Mvatas. 

B. Masculine: kdnikradatas (vdjrdt), ii.11.9. 

Genitive Singular Masculine and Neuter. 

A. Masculine. There are 127 oxytone forms (from 33 stems), 
and 53 barytone forms (from 34 stems). 

Oxytones : ardtiyatds ; satds, 10 ; yatds ; undatds ; rghdyatds, 3 ; rtd/yatds, 3 ; 
krnvatds, 2 ; karishyatds, 3 ; kshiyatds ; grnatds, 5 ; tujatds ; turatds ; twanyatds ; 
tudyatds ; dwhandyatds ; dvishatds, 2 ; dhrshatds, 2 ; prnatds ; pit&yatds ; brhatds, 
26; brahmanyatds, 2; mahatds, 11; mishatds ; minatds; rudhatds; vapushyatds; 
upatds ; vidhatas, 9 ; pucatds ; saparyatds ; sunvatds, 21 ; stuvatds, 9 ; huvatds. 

Barytones : dghnatas ; djiuryatas, 2 ; dbhufijatas ; drcatas, 2 ; dsunvatas, 3 ; 
inakshatas ; cdratas; cetatas ; jdratas, 2 ; jdyatas, 2 ; ji'vatas, 2 ; jyjydsatas; tu'i-v- 
atas; tr'shyatas; diddsatas; drdvatas, 2; dhdmatas ; dhrdjatas, 2; pi'yatas; 
bhu'shaias, 2 ; yudhyatas, 4 ; rebhatas, 2 ; ririkshatas ; rupatas ; vdyatas; vdrdh- 
atas; venatas; pdrdhatas, 3 ; (dcatas; si'datas; sishdsatas; jighdnsatas, 2 ; hdrsh- 
atas; and t'vatas, 3. 

In ii.27.3, ddabdhdso dipsanto, Roth conjectures dipsato : ' Un- 
deceived of him that would fain deceive.' 

Neuters : prdnatds, 2 ; satds, 4 ; krpayatds; brhatds, 5 ; mahat- 
ds; mishatds; ydtds; and, d'satas, p. &'-, - ; cdratas, 2 ; bhdvatas; 
rupatas; venatas; sishdsatas. 

B. Masculines : jigatas; tdritratas; dhdkshatas; dddatas, 2 ; 
dd'satas ; dddhatas, 4 ; dddhatas ; bdpsatas, 2 ; vdghdtas, 5 ; 
v&vishatas; pd'satas; jdnghnatas. 

Neuters: vd'vadatas; jdgatas, 29; dvdrvrtatas, AV. v. 1.8. 

Locative Singular Masculine and Neuter. 

A. Masculines: yati, 10; mahati; sunvati; and, irdyati; from 
AV., ubhayd'dati, v.31.3. Neuters : mahati, 2 ; and, dsati. 
At the beginning of i.l 13.10a and ii.30.1c?, we have with metrical 
lengthening, kiydti d', p. -&ti; see Rik Pr. ix.29. 

Neuter: jdgati, i.164.23. 



V.s.m.] Noun- Inflection in the Veda. 509 

Vocative Singular Masculine. 

A. Grassmann takes catdvan, vi.47.9, as a compound of dvant; 
but see stems in vant, V.s.m. The word bhos is explained as a 
contracted form of bhavas. If this be vocative of bhav-ant, we 
must regard it as formed after the analogy of the stems in vant, 
as though the stem were bha-vant. See stems in vant, V.s.m. 

Nominative, Accusative, and Vocative Dual Masculine. 

A. There are 79 d-forms (from 52 stems), and 13 aw-forms (from 

11 stems). 

I. The tf-forms: apndntd; sdntd, 4; ydntd, 4; gmdntd, 2 (and see below); gav- 
ydnid; dafasydnid, 3; dif&ntd; duhdntd; punantd; pricdntd ; hrhdntd, 4; bhuj- 
6/ntd ; mishdntd ; mindntd ; irajydntd ; vanvdntd ; tifdntd, 2 ; vasnayantd ; vdjay- 
antd; vidntd; vfdhdntd ; prnvantd; sapwyantd; svmnayantd; — dmardhantd; 
dyatantd ; drhantd ; ishdyantd ; ksh&yantd ; cdrcmtd, 2 ; coddycmtd, 2 ; jandyamtd ; 
jdra/ntd ; dr&oantd ; dhdmantd ; pdrdyantd, 2 ; pibantd ; mddantd, 4 ; rddantd ; 
vdjdyantd; vddantd; vdpantd; vdhantd, 6; venantd; sdda/iM; sdpantd; s&hantd; 
sd'dhantd; h&yantd;— pd'ntd; yd'nld; mahd'ntd, 3; brhantd, voc. 

II. The au-i orms : afnantau ; sdntau, 2 ; ydntau; ishdyantau ; Itrri'dantau, 2 ; 
kshdyantau ; tir&ntau ; ydjantau ; rd'jantav, ; vrdhantau ; and mahd'ntau. The 
AV has m&dantau, xviii. 1.54c, a slight modernization of B.V. x.14.'7c. 

Tn i. 122. 11, gm&ntd, the pada has gmdntd; cf. Prat, viii.15. 
The pada sometimes has a dual in -d against -& of the samhitd 
(p. 342) ; but not vice versa. The sense require* a plural ; see 
N.p.m, and Benfey, Abh. xix.260. 

B. Here belong: I. tdritratd; bdpsatd; bibhratd, 3; pd'satd; 
anupd'satd; — II. dddhatau; bibhratau, 2. 

Nominative and Accusative Dual Neuter. 

A. Here belong : yatt', 2 ; brhatf; and perhaps pr'shati, 
i.162.21— see p. 391. 

Dative and Genitive Dual Masculine. 

A. Datives: mahddbhydm; mrdayddbhydm, i.\SQ.\bis (Gr. 
-bhidm). Genitive: jdratos. Samyatds (accent), AV. xvi.8.22, 
is G.d.n. 

B. Genitive : pipratos. 

Nominative and Vocative Plural Masculine. 

A. There are 428 forms (from 166 stems). Examples are 
sdntas, 9; ydntas, 15; hrdntas; krnvdntas, 8; grndntas, 21 
devaydntas, 26; bhiddntas ; bhinddntas, 2; mddantas, 10 , 
ydjaydntas, 13 ; ghndntas, 9 ; ucantas, voc, x.30.2. Not 
included above are : satdmahdntas; mahd'ntas, 2 ; mahdntas, 
voc, 2. 

Gr. proposes gmdnto, text gmdntd, i.122.11 ; see N.d.m. For 
dipsantas, see G.s.m. 

We find ubhayd'ddtas, x.90.10; and ubhayatod&taa, Manu i.43. 
In each case, the absence of the nasal may be due to the metre, 
since the syllable dt is penultimate in an anvshtubh-pdda. The 
vol. x. 68 



510 C. R. Lanman, [Ptcp. in nt, I. 

AV. has the transition-form ddntds, iv.3.6 : v. 18.8, etc. ; similarly 
the Qat. B r-> -dantds (instead of -datis), i.6.3 30 . 
B. Here belong 53 forms (from 20 stems). 

Enumeration: kdrikratas ; jd'gratas ; jagatas, voc. ; titratas; dddatas,9; da'pa- 
tas, 2; dodhatas; dodhatas, 5; dt'dhiatas; ddvidhvatas, 3; nonuvaias; pipratas, 
2; bdpsatas, 2; bibhratas, 12; mdrmrjatas; vdgh&tas, 5; vdghatas, voc. ; pdpuca- 
tas; sisratas; j&hvatas ; jdhuvatas. 

For the irregular nasal of vavrdhdntas, iv.2.17 (=AV. xviii.3.22, 
vdv-), see p. 505. 

NOMINATIVE AND ACCUSATIVE PLURAL NEUTER. 

A. We have the forms mahd'nti, p. mahd'nti, 3 times, and 
AV. ix.5.1 : x.4.23 ; iydnti, p. iydnti, RV. vi.23.4 (Prat, ix.30) ; 
brha'nti, AV. viii.9.3. 

The only real participle in the N.A.p.n. is sd'nti, ii.28.1 : 
viii.8.23. The pada reads sd'nti; Prat, ix.25. 

In i,140.13cd, gdvyam ydvyarh ydnto dirghd' dhd[ni] \ tsham vdram amnio var- 
anta, Roth takes dhdni and arwiio as coordinate subjects of varanta. ' M6gen die 
lange sich dahinziehenden Tage und die Morgenrothen uns Vieh, Korn, und 
Lebensfrische schenken.' We might (1) read yd'nti; or (2) we might consider 
the masculine yantas as the general gender before two different ones following 
(n. and f.) ; or (3), in view of the extreme unfamiliarity of the neuter form, we 
may make the masc. do duty outright for the neuter, as in the case of haripriyam, 
p. 377. 

Accusative Plural Masculine. 

A. There are 65 oxytone forms (from 24 stems), and 24 bary- 
tone forms (from 15 stems). 

Oxytones: satas, 3; yatds, 4; rgftdyatas, 2; grnat&s, 1 ; tarushyatds ; tu&yatds, 
2; prnatds; pftanyatds, 9; brhatds, 3; bhandandyatds ; mandyatds ; mahatds, 2; 
rudaids; vanushyatas,6; upatas, 11; vdjayatds; vidhatds; pravasyatds ; prnvatds, 
2; prudhtyatds ; sakhtyatds ; sasatds,2; stuvat&s ; datds, vii.55.2 : AV. v.29.4: 
vi.56.3. 

Barytones : dprnatas, 2 ; dprayuchatas ; drcatas ; Jcshdyatas ; dhd'vatas ; dhraj- 
atas ; pdtatas, 2 ; rd'yatas ; rishatas ; r&rukshatas (ruh) ; vadatas ; vrd'dhatas, 5 ; 
pdrdhatas, 4; sisrpsatas ; i'vatas; pydvddatas, AV. xi.9.17. 

B. Here belong : jdkshatas (ghas) ; dd' satas ; pd'pvasatas; 
from AV., jd'gratas, vii.108.2 ; dodhatas, xii.1.58 : and so iii.6.2 ; 
pdpucatas, viii.3.13. For sapcatas, see pp. 469, 467. 

Instrumental Plural Masculine and Neuter. 

A. The oxytone stems (except ddnt) do not shift the accent to 
the ending. There are 33 masculine forms (from 21 stems), and 
7 neuter forms (from 5 stems). 

Masculines : dnirnishadbhis ; dprayucJiadbhis ; asredhadbhis ; citdyadbhis ; tujd- 
yadbhis ; devayddbhis, 2 ; dhd'vadbhis ; patdyadbhis ; brhddbhis, 3 ; mahddbhis, 
5 ; rebhadbkis ; rUpadbhis, 3 ; vddadbhis ; upddbhis ; vdjayddbhis ; vrdjadbhis, 2 ; 
pocadbhis; pucadbhis; pucdyadbhis, 2; pubhdyadbhis ; dadbhis, 2. 

Neuters : brhddbhis ; mahddbhis, 2 ; rupadbhis ; pucdyadbhis, 2 ; standyadbhis. 

Transition to the a-declension : ddntais, AV. xi.3.37. The 
Qat. Br. has dadbhis by the side of ddntds, iii.5.4". 



I. p.m. ii.] Noun- Inflection in the Veda. 511 

B. Here belong : nd'nadadbhis, 2 ; pdpruthadbhis ; vd'vadad- 
bhis; vdghddbhis, 4 ; $d' pvmadbhis. 

Dative and Ablative Plural Masculine. 

A. Datives : kshicdhyadbhyas ; tudyddbhyas ; brhddbhyas; 
mahddbhyas; vddadbhyas;- ugddbhyas; vdhadbhyas ; pdpyad- 
bhyas; prayddbhyas (Gr. -bhias), v.54.9 ; grnddbhyas (Gr. 
-bhias) ; sunvddbhias, i.132.4 ; dadbhyds, AV. xi.2.6. 

Ablative : pardyddbhyas, x.34.5. 

Genitive Plural Masculine and Neuter. 

A. Masculine. There are 50 oxytone forms (from 22 stems), 
and 10 barytone forms (from 7 stems). Here Gr. puts sthdtd'm, 
i.70.3 ; but see p. 423 med. 

Oxytones : acRivariyat&'m, 3 ; satd'm, 3 ; yatd'm ; rj&yatd'm ; grnatd'm, 3 ; 
jurat&'m ; devayatd'm, 4 ; dvishatd'rn ; brhatd'm ; mahatd'm, 5 ; ydtd'm, 13 ; 
varmshyatd'm; ufatd'm; fairAyatd'm, 3; pravasyatd'm ; prnvatd'm ; sakhtyatd'm ; 
sasatd'm; sunvatd'm; stuvatd'm; ghnatd'm ; vdjayatd'm (Gr. -dam), vi.45.29; 
sat&dm, vi.6l.l. The AV. has datd'm, v.23.3. 

Barytones: dmnvatdm; caraidm ; jayatdm, 4; nr'tyatdm; pdtatdm; vivdsatdm; 
(drdhatdm. 

Neuter : rathirdydtdm (accent! !), ix.93.4 ; see p. 506. 

B. Masculines: jdgatdm, 2; vdghdtdrn, 3; bibhyatdm, AV. 
viii.8.20. 

Locative Plural Masculine and Neuter. 

A. Masculines: grndtm, 2; devaydtm; patdyatsu; mahdtsu, 2. 
Neuter : jil'ryatsu. 

PARTICIPLES IN VAM 

For Brugman's view touching the original form of the vans 
and yans-stemB, see Zeitsch. xxiv.54. 

About 75 stems, simple and compound, end in vans. They 
have a three-fold form. In the strong cases the stem ends in 
vdns; in the weak, in ush; and in the middle, in vad (only three 
such forms in the Rik). They are declined only in the masculine 
and neuter; the feminine is formed by adding I to the weak 
stem (p. 367). The accent is always on the final syllable of the 
stem, save in negative compounds. 

Even in the Veda there are several instances in which the weak 
form of the stem seems to be used in strong cases : cakrdsham, 
emushdm, proshtisham, dbibhyushas (see A.s.m. and N.p.m.). It 
is possible to consider these as simple forms of transition to the 
a-declension (caler&sha-m, etc.). Indeed, we have one plain case 
of such transition in Cat. Br. xiv.l.2 u : td'm emilsha (Uf) Ui 
vardhd ajjaghdna. Of. Devamidhusha and Midhusha. 

In the Epos, similar anomalies are by no means rare. We 
have: divydstra-vidushau, N.d., MBh. iv.1847; vidushas, N.p., 
iii.15850 ; vedavidushas, Brhat Samhita, xvi.24. For other cita- 
tions, see BR. vi.1068; and cf. Kern, Translation of Brh. S., p. 100. 



512 C. R. Lanman, [Ptcp. in vans. 

Nominative Singular Masculine. 

The N.s.m. in -vdn occurs 240 times (from 44 stems). It appears 
in the sarhhitd as -vdnp once (iii.44.2, vidvd'hp c-, Prat, iv.32) ; cf., 
moreover, dcikitvdnp c-, AV. ix.9.7, with dcikitvdn c-, RV. i.164.6, 
and see Whitney to Ath. Pr. ii.27 end. It appears as -vdn, the 
regular elided form of -vdns (see p. 346), before a vowel, but only 
in the interior of a, pdda (Prat, iv.26). 

The form -vdfi occurs 70 times : before a-, 42 times, as v.2.8 ; before &-, in 
i.94.6: ii.6.8: iu.14.2: iv.8.4; 19.10; before i-, i.37.8; 174.6: v.29.3,14; 30.11; 
32.2: ix.106.6: x.112.5; 160.2; before t-, i.145.1 ; before a-, i.52.8 ; 190.7: iii. 
29.16; 35.4: x.11.6 ; before r-, i.146.5: iii.31.1 ; 60.3: viii.81.9 : x.2.1 ; 28.5; 
138.3 ; before e-, ix.96.2. 

Furthermore, -vM occurs before y- (ii.4.5 : ix.107.1), and v- (x.132.3). The 
Prat, enumerates all the cases at iv.28. The VS. reads <id y- (Pr&t. iii. 13 5) ! In 
ii.12.4:, jigvd'X (7?=nasalized I) occurs before I- ; Prat. iv.4. 

At the end of a pdda we have -vdn before a vowel (23 times) : so before a-, 
e. g. vii.7.5 ; before d-, e. g. i.169.3 ; before i-, e. g. x.8.8 ; before «-, e. g. iv.7.8 ; 
before r-, x. 148.3. In three exceptional cases we have -vdfl, even at the end of a 
pdda : in i.69.6c and iii.32.6a (see Pritt. iv.26), and in i.69.1a (aksharapahkti). 

Enumeration : dcikitvdn ; aproshivdn ; drarivdn ; avidvdn, 3 ; cakrvd'n, 2 ; 
jaganvd'n, 6 ; cikitvd'n, 32 ; jigivd'n, 2 ; jvjjurvd'n, 3 ; jujushvd'n, 2 ; jAjuvd'n ; 
jajnivd'n ; tatanvd'n, 2 ; dadacvd'n ; dadasvd'n ; dadvd'n ; ddcvd'n, 5 ; dtdivd'n ; 
dadrcvd'n; dadhanvd'n, 3; dadhrshvd'n, 3 ; papivd'n, 8; pvpushvd'n; paprivd'n, 
2; bibhivd'n; babk&vd'n, 3 ; mtdhvd'n, 5 ; mamrvd'n; yayivd'n; rarivd'n ; rurvk- 
vd'n ; vavanvd'n ; vivikvd'n ; vividvd'n, 2 ; vidvd'n, 98 : vividhvd'n ; cucukvd'n, 3 ; 
cucruvd'n; sasavd'n, 6; sdsdhvd'n, 2; sdhvd'n, 9; iasthivd'n; jaghanvd'n, 17. 
Here Gr. puts cvdstvdn — BR. stem cvasivant. 

Nominative and Accusative Singular Neuter. 

Here belong tatanvdt (vi.21.3) and samvavrtvdt (v.31.3) — both 
accusative. 

Accusative Singular Masculine. 

Here belong 66 forms (from 27 stems). 

Enumeration: tyivd'nsam, 2; cakrvd'nsam ; jdgrvd'nsam ; cakhvd'nsam; cikit- 
vd'nsam; jUjuvd'nsam ; ddcvd'nsam, 9; didivd'nsam, 11; dti/rvidvdnsam ; papti- 
vd'nsam ; papivd'nsam, 3 ; ptpivd'nsam, 2 ; paprivd'nsam ; mamrvd'nsam ; 
rvrihvd'nsam; vidvd'nsam; vavrivd'nsam, 5 ; vdvrdhvd'nsam, 2 ; cAcuvd'nsam ; 
sasavd'nsam ; sdsahvd'nsam, 2 ; mshuvd'nsam ; s&vidvdnsam ; sasrvd'nsam ; 
tastabhvd'nsam ; tasthivd'nsam, 7 : sushupvd'nsam. 

There are several remarkable instances in which the weak form 
seems to be used for the strong. They are : cakrtisham (for 
cakrvd'nsam — Say., krtavantam), x.137.1 (=AV. iv.13.1) ; vard- 
hdm emushdm (for emivd'nsani), viii.66.I0 ; pitdram prosh'usham 
(for proshivd'nsam), Qat.Br. x ii.5.2 8 . The accent of etnushd is 
perhaps due to a misconception of the form (BR.). 

Instrumental Singular Masculine and Neuter. 

Masculines: dbibhyushd (i.6.7) ; cikittishd; vidti$hd,3. Neu- 
ters : dbibhyushd; bibhyushd. 

Dative Singular Masculine. 

Here belong : drarushe; flcdshe, 2 ; cakrdshe; jagm&she; cikit- 
tishe,6; jigytishe, 2; daddptishe, 4; ddp&she, 111; bibhytishe; 
midhtishe, 11 ; vid&she, 5 ; sedtishe. 



Ab.Gr.s.m.n.] Noun- Inflection in the Veda. 513 

Ablative and Genitive Singular Masculine and Neuter. 

Ablatives : (masc.) drarushas; jujurushas, 2 ; (neut.) tastlv&shas 
(1.6.1?). 

Genitives : (masc. — 63 forms, from 1 9 stems) dddpushas, 2 ; 
drarushas, 6 ; iyushas / jagmtishas / cikitfishas / jigydshas, 4 ; 
jdnushas; tatardshas; dadtishas, 2 ; ddpiishas, 21 ; didiykshas ; 
paptishas; bibhy&shds; rnidk&shas, 9 ; vividiishas; sedtishas, 2 ; 
sushuv&shas; tasth&shas, 6 ; jaghntishas; mamrdshas, AV. vi. 
18.2; (neut.) vavavrushas. 

Delbrtick, Verbum, p. 235, says that dadhanvdtas, vi.48.18, may 
be G.s.m. of da-dhan-vdns, for dardhanv-ush-as; cf. p. 516. 

Vocative Singular Masculine. 

Here belong: cikitvas, 11 ; titirvas; didivas, 6 ; midhvas, 10. 
The AV., at vii.97.1 (=RV. iii.29.16), has cikitvan, a moderniza- 
tion of the Rik-form cikitvas; cf. vant-stems, V.s.m. 

Nominative and Accusative Dual Masculine. 

I. The dr forms: okivd'nsd; jaganvd'hsd; jdgrvd'nsd; didi- 
vd'nsd; papivd'hsd; vavanvd'nsd ; vidvd'nsd, 3; pdpuvd'nsd ; 
pupruvd'nsd; tasthivd'nsd, 2. II. There is but one aw-form, 
vidvd'nsau. 

Nominative Plural Masculine. 

Here belong 43 forms (from 22 stems). 

Enumeration : avidvdnsas ; cakrvd'nsas, 2 ; jaganvd'iisas ; jdgrvd'nsas, 3 ; cikit- 
vd'nsas ; jigtvd'nsas, 4 ; titirvd'iisas ; dadrvd'iisas ; ddfvd'nsas, 2 ; paptivd'nsas ; 
mtdhvd'nsas; ririkvd'nsas, 2; vidvd'nsas, 3; fv^ukvd'nsas; f&puvd'nsas ; sasa- 
vd'nsas, 5 ; sdsahvd'nsas ; sdhvd'nsas ; sasrvd'nsas, 2 ; tushtuvd'nsas, 3 ; tasthi- 
vd'nsas, 5 ; sushupvd'iisas. For bhaktivd'nsas, AV. vi.79.3 (cf. VU.9Y.3), see BR 
v.168. 

Weak form for strong. In i.11.5 (=SV. ii.601), dbibhyushas 
can hardly be construed otherwise than as N.p.m. (for dbibhivdn- 
sas) with devd's. Mtlller, Translation, p. 3, takes pdri tasthtishas, 
i.6.1, as equivalent to paritasthivd'nsas ; cf. Mahidhara to VS. 
xxiii.5 : tasthushah vibhakter vyatyayah tasthivdnsah karmdrtharh 
sthitd rtvijah; but the antithesis seems to be between pdri car 
and sthd. See A.p.m. Under this category Bollensen, Z.D.M. G. 
xxii.604, puts cikifdshas (i.T3.1) and daddshas (i.54.8 — Say., 
havirdattavantas .... yad vd . . .). Both forms are genitives 
s. m. 

Nominative and Accusative Plural Neuter. 

There seems to be no example of a form in -vdnsi. 

Accusative Plural Masculine. 

Here belong : jagmtishas; cikifushas, 2 ; jigyiishas, 2 ; ddp- 
■dshas; midh-dshas, 3 ; vid'ushas; sedushas; tasthdshas, v.53.2 : 
ix.19.6 : and i.6.1 (A.p.m. rather than Ab.s.n., on account of 
pdri). 



514 0. R. Lanman, [yans-stems. 

Instrumental and Genitive Pltjbal Masculine. 

Instrumental : jdgrvddbhis, 3. 

Genitives : dddpushdm; jigydshdm; dad'&shdm; midh'dshdm; 
vid-Ctshdm; pupruwtishdm, Qat. Br. ii.2.2". 



COMPARATIVES IN YA&S. 

These stems have the form -ydns in the strong cases, and -yas 
in the weak. No middle form occurs. They are declined only in 
the masculine and neuter; the feminine is formed by adding I to 
the weak stem (p. 367). We find tdviyans and tdvyans, ndviyans 
and ndvyans, pdniyans and pdnyans, sdhiyans and sdhyans, side 
by side. 

Benfey, Gram. p. 310, note 5, mentions kaniyas-am as a strong 
case-form from a weak stem (for kaniydns-am). It occurs thrice 
in the MBh. Like it is kaniyasau. BR. refer both forms to an 
o-stem (Jcaniyasa-m). Both are at all events by origin transition- 
forms. Compare the stems preydnsa, preyasa, and compounds. 

Nominative Singular Masculine. 

The N.s.m. in -dn occurs 56 times (from 20 stems). It appears 
in the samhitd as -dnp once (x.90.3, jyd'ydnp ea) ; cf. the sandhi 
of the AV. at xiii.4.46,47. For the rest, it is treated just as the 
N.s.m. of -vans, and appears as -dn before vowels in the interior 
of a pdda four times (before a-, iv.30.1 : vii.98.1 ; before i-, v.41.12 : 
viii.1.6) ; but as -dn before vowels at the end of a, pdda nine times 
(e. g., before r-, i.110.7 : vi.12.5). 

Enumeration: dtavydn, 2; djiydn, 2 ; kaniydn ; jdvtydn, 5 ; jyd'ydn, 7 ; tariydn; 
tavtydn,3; tavydn,3; dhdviydn; naviydn,3; manhiydn; y djiydn, 15 ; yddhiydn; 
vdniydn, 2 ; variydn ; vasydn, 2 ; vedtydn ; preydn, 2 ; sahtydn, 2 ; skdbMydn. 

Nominative and Accusative Singular Neuter. 

Here belong 82 forms (from 16 stems) : r'jiyas; djiyas; kdniyas, 
2; jyd'yas, 2; tdviyas; ddviyas; drd'ghiyas, 5; ndviyas, 12 ; 
ndvyas, 7 ; n'ediyas, 3 ; preyas; bhU'yas, 7 ; bhtfyas-bhtiyas; 
vdriyas, 14 ; vdrshiyas; vdsyas, 20 ; svd'diyas, 3. 

In i.]Q5.l5d, the metre calls for the longer grammatical form 
ndviyas instead of the shorter equivalent ndvyas; similarly, in 
vi.30.la, for bhdviyas (not bhtl'ias — cf. i.83.1), instead of bhilyas. 
For ndv[i]yas in i.61.135, see p. 338 med. 

Accusative Singular Masculine. 

Here belong: jyd'ydnsam; tdvydnsam; drd'ghiydhsam; ndv- 
ydnsam; pdnydnsam ; pdpiydnsam ; preydnsam ; from AV., 
vdrshiydnsam,' ix.6.19 : xv.11.5 ; sdhiydnsam, xvii.1-5. For 
kaniyas-am, see above. 



I.s.m.n.] Noun- Inflection in the Veda. 515 

Instrumental Singular Masculine and Neuter. 

Masculines: jdviyasd, 2; ndvyasd, vi.6.1; bhil'yasd. 

Neuters : tejiyasd; tvdkshiyasd; ndviyasd, 3 ; ndvyasd, 4 ; 
pdnyasd,2; bhdviyasd; bhil'yasd (iv.24:.9bis) ; vdsyasd; vdsyasd- 
vasyasd; sdhiyasd, 2. 

Dative Singular Masculine and Neuter. 

Masculines : tdvyase; ndviyase; pdniyase; pdnyase; sdnyase; 
sdhiyase (read sdhyase, i.71.4c) ; sdhyase; bdliyase, AV. iii.29.3. 
Neuters: ndviyase; ndvyase, 10; sdnyase, 2. 

Ablative and Genitive Singular Masculine and Neuter. 

Ablatives : (masc.) tdviyasas; rabhyasas; sdhiyasas, 2 ; sdh- 
yasas, 2 ; (neut.) bhd'yasas. 

Genitives : (masc.) kdniyasas, 2 ; jyd'yasas; tdvyasas; ndvt- 
yasas; ndvyasas; bhU'yasas; (neut.) ndvyasas, 2. 

Locative and Tocative Singular Masculine. 
Locative : sdhiyasi, 2. Vocatives : .Sjiyas; jyd'yas. 

Nominative Plural Masculine. 

Here belong: $r'eydhsas; from AV., ti'kshniydnsas, iii.19.4Sw/ 
bhd'ydnsas, vii.60.7. 

Nominative and Accusative Plural Neuter, 
Here belongs ndvydnsi, nom., i.38.3a (catalectic ?). 

Accusative Plural Masculine. 

Here belong: kdniyasas; n'ediyasas, 2; bhil'yasas, 4; vdsyasas, 
14; vdhiyasas; vdrshiyasas, AV. vi. 136.2. 

Genitive Plural Masculine. 
Instead of the masc. ndvyasdm, we find tbe fem. form ndvyas- 
indm (martitdm) at the end of v.53.105 (7 syll., catalectic — Gr. 
-aam), and of v.58.15 (11 syll). The metre is chargeable with 
the irregularity. 

STEMS IN VANT AND MANT. 

The stems formed by the suffixes vant and mant are declined 
entirely alike, and it has therefore seemed best not to treat them 
in separate sections. The forms from stems in vant are given 
case by case under category A ; and those from stems in mant, 
under category B. 

These stems are declined only in the masculine and neuter. 
The feminine is formed by adding i to the weak stem (p. 367). 
The stem in vant, mant seems, however, to be used sometimes as 
a feminine ; thus, according to Gr., we have : pdpvatdtdnd, i.26.6 : 
ix.1.6 (I have taken the two words adverbially — see p. 480); 



516 C. R. Lanman, [vant, mant-stems. 

havishmatd devdtdtd, i.l28.2c/ rdthavate prdpastaye, \.122.\ld 
(the metre demands -vatyai). It may be questioned whether these 
are syntactical or formal peculiarities ; or due to a simple disre- 
gard of grammatical gender (cf. brhat'e, ix.15.2; 96.4: AV. 
xiv.2.72 ; ndmdbibhrati' dtithis, xv.13.6). 

These stems show the distinction of strong and weak in the 
retention of the nasal in the strong cases and its rejection in the 
weak. Several instances of the confusion of this distinction are 
probable. 

Weak form for strong. The most probable instance seems to 
me to be kshumdti ytithd' (for kshumdnti, A.p.n.), iv.2.18. Not 
improbable is x.59.1, sthd'tdreva krdtumatd rdthasya (for krdtu- 
mantd, N.d.m.) : ' Discerning are the two guides as it were of 
the wagon;' figuratively for the eyes of the body. Gr. makes 
indrdvato a N.p. in iv.27.4, rjipyd im 'indrdvato nd bhujy'&m \ 
pyend jabhdra; see Ueb. i.134. 

Strong for weak. There is no certain instance. The possible 
or supposed ones — prdyasvantas, havishmantas, vadMb'mantas, 
dyumantds — are discussed below ; see A.p.m. Kuhn, Beitrage, 
iii.475, proposes vd'java\n\tas in vi.50.Ua, taking it, I suppose, 
as G.s.m. This form is certainly demanded by the cadence, and 
perhaps it once stood in the text as N.p.m. with te, being dis- 
placed by the formal parallelism of the genitives. Rather than 
rdthava[n]te, i.122.11, read -vatyai. 

Unlike participles, these stems do not shift the accent to the 
ending when oxytone ; e. g. nrvdtd, revdtas — cf. jdnate, tujatds. 

Transition to the a-declension. I have met with no instance of 
this from vant or mant-stems; but cf. the Pali, himavanto va 
pabbato (parvato), Dhammapada 304, and Kuhn, Pdli-gram. p. 76. 

Transitions (?) from the aw-declension. The stem maghdvan has 
several supplementary middle forms from maghdvant in the Rik : 
maghdvdn, once ; -vadbhis, once ; -vadbhyas, 19 ; -vatsu, 8. So 
sahd'van makes sahd'vdn, 2. 

Similarly the stem yuvan has the supplementary form y&vat 
(A.s.n., i. 111.1 : x.39.8), both times in places where the regular 
form yuvd would be excluded by the metre : e. g. tdkshan 
pitr'bhydm rbhdvo yitvad vdyah. Cf. aw-stems, N.A.s.n. So in 
Suparnadhyaya, vii.4 (Ind. Stud, xiv.6), occurs yuvdtas, A.p.m. 

The I.s.m. varimdtd, i.108.2, is rightly referred by BR., vii. 
1800, to varimdn. The Sanskrit form varimnd' would not fill 
out the verse. Cf. da-dhan-vdt-as, p. 513. 

Some of these supplementary forms are probably forms of tran- 
sition to the «aw<-declension. For some, the N.s.m. may have 
served as the point of departure. Cf. maghdvd with maghdvdn 
r-; svdvdn y- (i.35.10) and dadhanvd'n y- (ix.107.1) with -vd y- of 
the VS. (xxxiv.26 : xix.2) — Pr. iii.135. See also Zeitsch. xxiv.53. 

Conversely, the stem drvant has two forms from drvan : drvd, 
21 ; drvdnam. 



N.s.m.] Noun- Inflection in the Veda. 517 

Nominative Singular Masculine. 

A. The N.8.m. in -vdn occurs 192 times (from 89 stems). 

B. The N.s.m. in -mdn occurs 100 times (from 29 stems). 
These forms are treated in the samhitd quite like those from 

stems in -vdns; see p. 512. 

The form ■vdfi, occurs in the interior of a pdda 50 times : before a-, 22 times, as 
v.28.4 (and at the end of i.70.5a, dksharapankti) ; before i-, 7 times, as iv.4.1 ; 
before U-, 11 times, as vi.47.1 ; before r-, i.189.6: ii.25.3: iv.16.1; 26.1: x.64.16; 
before d-, i.84.9: viii.86.4; before e-, iv.2.5: x.3.7 ; before o-, i.173.6. At the 
end of a pdda we have -vdn before vowels, 8 times, as iii.39.4: vi.37.1 ; and 
exceptionally in the interior of a pdda (Pr&t. iv.31) at i.23.23 : iv.16.9: x.9.9 ; 
90.3. The sandhi is -vdn before c-, j-, (- (ch-) ; e. g. vii.20.1 : ix.90.3: i.95.1: and 
-vdn before t-; as i.30.14. 

In like manner, -mdn" occurs in the interior of a pdda 47 times : before a-, 22 
times, as iv.2.5; 41.1; before i-, 12 times, as ix.2.9; before «-, 4 times, as 
vii.71.4; before r-, vi.18.2 : ix.96.13 ; 97.48; 110.11: x.89.5 ; before d-, i.12.9: 
ix.44.4 : x.98.3 ; before #-, ix.68.8. At the end of a pdda, we have -mdn before 
vowels in ii. 12. 12: vii.85.4: x.8.2. We find -mdn before c-, i.55.1. 

The fact that arvdn stands in the samhitd as arvdh" (i.163.13, before a-; 
ix.97.25, before i-) does not seem to me to justify the assumption that arvd is the 
nominative to arvant (Miiller, Translation, p. 56). The sandhi is perfectly regular. 
"We have, to be sure, the supplementary nom. arvd ; so maghavdn and maghdvd 
sahd'vdn and sahd'vd. 

A. The forms are : akshanvd'n ; agnivd'n ; angirasvdn, 2 ; Annavdn ; Apavt/ra- 
vdn ; dmavdn, 4 ; arvdn, 2 ; dft'rvdn ; iddvdn ; ishavdn ; r'ghdvdn, 2 ; etd'vdn 
kakshi'vdn, 4 ; kshapd'vdn, 2 ; kshdpdvdn, 3 ; ghr'nivdn ; janivdn ; tdpasvdn 
tarshid'vdn ; tdvasvdn ; t&visMvdn ; td'vdn ; tuvtrdvdn, 2 ; tvd'vdn, 9 ; dans&nd 
van, 2 ; dakshindvdn, 2 ; dakshindvdn, 2 ; datravdn ; dd'navdn ; dd'svdn, 3 ; dur 
hdndvdn, 2; devavdn; devd'vdn ; dyumnavdn ; n&masvdn, 6; niyidvdn, 12 
nt'lavdn ; p&yasvdn, 2 ; pavitravdn ; paviravdn ; pastidvdn ; puramdhivdn 
p&shanvd'n; praja'vdn, 4; prayasvdn, 6; pravatedn; prahd'vdn; bh&gavdn, 2 
maghavdn; matdvdn; matsardvdn; manasvdn; mar&tvdn, 12 ; mdhasvdn; mdyd'- 
vdn; md'hindvdn, 2; mehdndvdn; y&fasvdn; ydlumd'vdn; rdbhasvdn; rasavdn, 
revd'n, 10 ; vd'javdn ; vdjintvdn, 3 ; vivakvd'n ; vivasvdn ; vr'shanvdn, 3 ; vr'sh 
nidvdn ; factvdn, 2 j faradvdn ; p&pvdn ; fipravdn ; fiprinivdn ; (Amtvdn, 5 ; sakhi- 
vdn ; saihd'vdn ; sarasvdn ; sahasvdn, 3 ; sahd'vdn, 2 ; sahdvdn, 2 ; sutd'vdn, 2 ; 
s&nrtd'vdn ; stavd'n, 3 ; svadhd'vdn, 9 ; svddhittvdn ; s&arvdn, 3 ; hdrivdn, 4 
hitd'vdn; hemid 'vdn. 

B. The forms are: abdimd'n; dvimdn ; afdnimdn ; ishwmdn; udanimd'n 
rbhumd% 2 ; kak&dmdn, 2 ; krdtumdn, 4 ; garutmdn, 2 ; gomdn, 3 ; t&vishmdn, 
12 ; tvashtrmdn ; tvisMmdn ; dyumd'n, 11 ; dhrajimdn, 2 ; nadanumdfn ; parafu- 
md'n; papamd'n; pitumd'n, 4 ; barhishmdn ; mddhumdn, 20; vdswmdn; virtik- 
mdn; vrshtima'n, 2; farumdn ; focishmdn; sushumd'n; havishmdn, 19 ; hirimdn. 

It seems necessary to read i.ll0.9£ with elision and crasis 
rbhum'endra citrdm d' darshi rd'dhah, text -md'n indra. 

The resolution havishmaan, i. 127. 10^, is inadmissible; the 
pdda is catalectic (7 syll.). 

If we refer tuvtrdvdn, with BR, to tuvirdvant, for tuvi-rava- 
vant, the shortened form would be paralleled by kdnikrat etc. 
(p. 505) ; but see as-stems, N.s.m. 

Nominative and Accusative Singular Neuter. 

There are 204 forms in -vat (from 51 stems), and 81 forms in 
-mat (from 21 stems). 

VOL. X. 69 



518 C. R. Lanman, [vant, mant-stems. 

A. The forms are : apdsMhdvat ; dmavat, 6 ; arcivdi, 2 ; dfvavat, 2 ; dpvdvat, 
14; dtmanvdt; dvr'tvat; irdvat, 2; r'ghdvat ; etd'vat, 5 ; kshaitavat; gopd'vat 
grdbhanavat ; ghrtdvat, 9 ; candrdvat ; td'vat, 4 ; tokdvat ; tvd'vat ; d&kshindvat 
dyumndvat, 2; dhdnd'vat; dhvasmanvdt; ndmasvat; ni'lavat; nrvdt, 12; padvdi. 
2; p&yasvat, 2; pasiidvat; pr'shadvat; prajd'vat, 10; JorAa«4t'a< ; yd' vat, 1 
y&vat, 2 (p. 516); rafiwitia*. 3; rdsavat; revdt, 23; vayunavai; vayHndvat; vd'ja 
vat, 4 ; jjipr 'kvat ; vivdsvat ; vishdvat ; virdvai, 1 8 ; fatdvat, 4 ; faphdvat, 2 
fdfvat, 22; popwrf-pafDai ; sahdsravat, 3; sdhasvat; sinavat; suarvat, 6; hiranya 
vat, 11. 

B. The forms are : abhishtimdt ; rbhumdt ; Jcetumdt ; kratumat ; kshumat ; 
gomat, 23; jyotishmat; tvishtmat; dasmdt; da'numat ; dyv/mdt, 16 ; nidhim&t; 
pafwmdt, 2 ; pitumdt ; pushtimdt, 2 ; mddhumat, 18 ; manyumdt ; y&vamat, 3 ; 
rayimdt; vasv/mat, 2; suastimat 2. 

BR. take gnd'vas as N.s.n. with sajdtiam, ii. 1.5, for gnd'vat; 
Weber would read gndvas, as voc. 

Accusative Singular Masculine. 

There are 86 forms in -vantam (from 43 stems), and 93 in 
-mantam (from 19 stems). 

A. The forms are : apidhd'navantam ; apUpdvantam, 2 ; anantam, 1 ; d^vd- 
vantam, 4 ; asihanvdntam, ; dtmanvdniam ; indrrasvantam ; ifrjasvantam ; ti'rnd- 
vantam ; etd' 'vantam ; enasvantam ; ojasvantam ; omanvantam ; amid' vantam ; 
kaksht'vantam, 4 ; ghrtdvantam, 8 ; dd'svantam ; devdvantam, 2 : dhdnd'vantam, 

2 ; dhf'vantam ; nrvdntam ; padvantam ; parasvantam ; prajd' vantam, 3 ; bhd's- 
vantam ; marutvantam, 1 ; mahishvantam ; rdtnavantam ; revantam, 2 ; vapd'- 
vantam, 2 ; vayd'vantam ; vd'javantam, 2 ; vd'ravantam ; vivasvantam ; vira- 
vantam, 1 ; vr'shanvantam ; (atdvantam, 2 ; fdpvantam, 2 ; sacand'vantam ; 
sdrasvantam, 2 ; harivantam, 2 ; hdstavantom ; hiranyavantam. 

B. The forms are : rbhumdntam ; kdnvamantam ; ketwndntam ; kshumAntam, 
5 ; gmnantam, 23 ; jyotishmantam, 2 ; tdvishimantam ; dyuma/ntam, 16 ; nidhi- 
mantam ; pitumdntam ; bdndhumantam ; bhdnumdntam ; mddhumantam, 25 ; 
vdsumantam, 1 ; vd'tfmantam ; vrshtimdntam ; (rushtimdntam, 2 ; liavfahmantam, 
2 ; hotrmantam. 

Instrumental Singular Masculine and Neuter. 

A. There are 25 masculines (from 14 stems), and 10 neuters 
(from 8 stems). 

B. There are 14 masculines (from 9 stems), and 5 neuters (from 

3 stems). 

A. Masculines: drvatd, 10; dpvdvatd ; udanvdtd ; r'kvatd; kakshi'vatd ; niyut- 
vatd,2; nrvdtd; rnar&tvatd, 2 ; ydpasvata; revdtd; vipvadeoidvatd ; vish&vdtd ; 
fubhrd'vatd ; sdhasvatd. 

Neuters: etd'vatd; candrdvatd; prajd' vata, 2 ; barhdndvatd; vivdsvatd; s&nrtd'- 
vatd; hdritvatd; fdpvatd, 2. 

B. Masculines : gdmatd ; jyotishmatd ; divitmatd ; dyumdtd ; bhrshtimdtd ; 
varimdtd (see p. 516) ; vasumatd, 4 ; virukmatd, 2 ; havishmatd, 2. 

Neuters: divitmatd; virukmatd; havishmatd, 3. 

In i.26.6 and ix.1.6, I take fdpvatd as neuter, adverbially — not 
with tdnd, which as a noun would be fern. In i,128.2c, however, 
havishmatd seems to go with the fem. devdtdtd; see p. 516 top. 

Dative Singular Masculine and Neuter. 
A. There are 51 masculines (from 21 stems), and 4 neuters 
(from 3 stems). 



D.s.m.n.] Noun- Inflection in the Veda. 519 

B. There are 13 masculines (from 7 stems), and no neuters. 

A. Masculines: drvate, 10; dpvdvate; kaksM'vate, 3; ddk- 
shindvate; ddkshindvate; datvdte; dd'svate,2; niyutvate {\. 185.1); 
pUshanvdte, 2 ; marutvate, 8 ; md'vate, 6 ; yuvd'vate; rathavate; 
vivdsvate, 2 ; vivasvate ; pdctvate ; pdpvate, 3 ; sdhasvate, 3 ; 
silnrtd'vate; suarvate; hdrivate. 

Neuters : padvdte, 2 ; revdte; pdpvate. 

B. Masculines : gdmate / cdkshushmate ; tvishimate / div'it- 
mate; dyumdte; barhishmate, 3 ; havishmate, 5. 

In i.l22.11c?, both grammar and metre favor the reading 
rdthavatyai ; and this I would suggest in place of rdthavate 
(prdpastaye). In vi.17.14, Gr. reads dyumdte i-, p. -tah i-. 

Ablative Singular Masculine. 

A. Here belong : tud'vatas (viii.45.35) ; vivdsvatas; vr'shnid- 
vatas. 

Genitive Singular Masculine and Neuter. 

A. There are 88 masculines (from 30 stems), and 6 neuters 
(from 4 stems). 

B. There are 24 masculines (from 4 stems), and one neuter. 

A. Masculines : arvatas, 9 ; dpvdvatas, 2 ; kaksht'vatas ; jd'vatas ; tvd'vatas, 8 ; 
dadhanvatas ; dd'svatas; devdvatas, 2; dht'vatas ; dht'vatas-dhivatas • niy&tvatas ; 
nrvatas, 2; prajd'vatas, 4; marutvatas, 2; md'vatas, 3; mehan&vatas ; ydfasvatas, 
2; revatas, 1 ; vayasvatas, 2 ; vd'javatas; vivasvatas, 19 ; vivasvatas, 2 ; viravatas; 
fdcivatas ; f&pvatas, 3 ; sdrasvatas ; sahasvatas, 2 ; sutd'vatas, 4 ; swrdvatas ; 
svadhd'vatas ; hr'shivatas. 

Neuters : etd'vatas, 3 ; yd'vatas ; vish&vdtas ; fafvatas. 

B. Masculines: kshumdtas, 2; gdmatas, 19; dyumdtas; havishmatas, 2. Neu- 
ter: gomatas. 

For dadhanvatas, cf. p. 513. 

Locative Singular Masculine and Neuter. 

A. There are 18 masculines (from 9 stems): drvati, 2; ddk- 
shindvati; nrvdti; pastidvati; ydpasvati; vdnanvati, 2; vivds- 
vati, 3 ; vivasvati; paryand'vati, 6. 

Neuters: dpvdvati; ptmivati; suarvati. 

B. There is but one example, gdmati, masc, 8. 

Vocative Singular Masculine. 

I. In the Rik, the vocative of stems in -vant and -mant ends in 
-vas and -mas — and not, as in Sanskrit, in -van and -man. The 
like ending in -as is seen elsewhere only in the vocatives cikitvas, 
titirvas, didivas, midhvas (p. 513), ojiyas, jydyas (p. 515), rtdvas, 
evaydvas, prdtaritvas, mdtaripvas, khidvas, and pumas. 

A. Here belong 107 forms (from 14 stems) : rsMvas; gndvas; 
tavishivas, 2 ; niyutvas,2; marutvas,3; rayivas,3; vajrivas, 15; 
vtravas; paktivas; paeivas, 13; sarasvas; sahasvas, 7; svadhdvas, 
19; harivas, 38. Here Gr. puts vibhdvas, i.58.9; it ought to be 
referred to vibhd'van. For gnd'vas, ii. 1.5, see A.s.n. In x.74.5, 
BR. read pdci vas for pdoivas. 



520 0. R. Lanman, \yant, mant-stems. 

In the following repetitions of Rik-passages, we find the form 
-vas: VS. xxvi.21 (RV. i.15.3) : vii.35 (iii.51.7) : xviii.74 (vi.5.7). 
It is interesting to see how the Samaveda, at i.351, in its rendering 
of RV. vi.44.1, avoids the antique and no longer understood form 
(yd) rayivo — namely, by the bad variant, (yd) rayim vo. In 
genuine Atharvan verses, there is not a single vocative in -vas or 
-mas: svadhdvas occurs in a Rik-passage (xviii.l.26=RV. x.11.8) : 
and for bhagavas, xix.34.8 (??), the ed. reads samdbhavas. The 
form bhagavas occurs VS. xvi.52-3: TS. iv.1.10": Ait. Br. iii.20: 
v.14: viii.24; bhagos, (>t.Br. xiv.5.4': 7.3 s . Cf. bhos, p. 509; 
and aghos (aghavas). 

B. Here belqng 8 forms (from 6 stems) : tuvishmas; dyumas; 
bhdnumas; $uci$hmas; havishmas; mantumas, i.42.5 : vi.56.4 : 
x.!34.6=SV. ii.441. 

II. A. The modern form in -van occurs also sporadically in the 
Rik as follows: arvan, i. 163. 1,3,4,8,11 (a well-known late hymn): 
vi.12.6; pavasdvan, i.62.11; patdvan, vi.47.9. Since the pada 
reads patd-van (cf. Prat, ix.10), it is better to take this word from 
patdrvant, and not, with Gr., from $ata-\-avant. 

Excepting x.11.8 (above), the only Rik-verses containing a 
vocative in -vas which are repeated in the Atharvan are RV. v.42.4 
and x.84.1 ; and in both, the later text modernizes the form to 
-van (cf. p. 513), and reads harivan (AV. vii.97.2) and rnarutvan 
(AV. iv.31.1). Besides these the AV. has : vdjinivan, iv.38.6,7 ; 
vrshnydvan, v.25.8; svadhdvan, v.ll. 4,5,11. 

B. There is no example of a voc. in -man. 

Nominative, Accusative, and Vocative Dual Masculine. 

I. A. Here belong 20 forms (from 18 stems) : dngirasvantd; 
dnasvantd; drvantd, 2; indravantd; kdpdvantd; Mpavantd; 
dhdrrnavantd; ndmasvantd; niytitvantd; pavitravantd; mardt- 
vantd; mitrd'-vdrunavantd; vd'javantd; vishnuvantd; vydcas- 
vantd, 2 ; pdpvantd; pd'tavantd ? ; sdptlvantd. 

B. Here belong : arcimdntd; rbhumdntd; krdtumantd; va- 
dhUlmantd. 

n. A. Here belong : ashthivdntau; rdmanvantati; and from 
AV., ashthivdntau, 5 ; yd'vantau, xii.3.1 ; sdhasvantau, xix.32.5 ; 
svddhdvantau, voc, v. 9.8. 

Weak form for strong : krdtumatd, x.59. L ; see p. 516. 

Ablative and Genitive Dual Masculine. 
A. Ablative : ashthivddbhydm. Genitives : vdjinivatos ; 

f&mtvatos; sdrasvativatos. 

Nominative and Vocative Plueal Masculine. 

A. There are 91 forms (from 46 stems). 

B. There are 53 forms (from 17 stems). 

A. The forms are : akshanvdntas ; anasvantas ; amavantas, 3 ; Arvantas, 11 ; 
dpodvantas ; dfi'rvantas ; iddvantas ; indravantas, t ; A'rjasvantas ; enasvamtas ; 
kaksht'vantas ; karnavantas ; ghrtavantas, 2; cashd'lmtmtas ; ddkshindvantas, 2 j 



N.V.p.m.] Noun- Inflection in the Veda. 521 

dd'manvantas, 2; dravinasvantas ; dht'vantas; namasvantas ; niyutvantas ; nrvan- 
tas ; patnivantas, 3 ; padv&ntas ; pavitravantas, 2 ; pd'jasvantas ; piisktd'vantas ; 
pUshanvantas ; prajd'vantas, 2; prdyasvantas, 10; pravdtvantas ; bhagavantas, 3 ; 
marutoantas ; yajnavantas; vacand'vantas ; vwrmanvantas; viravantas; vrci'vantas ; 
paktivantas ; papvantas, 4; pimtvantas ; sdptivantas ; sutd'vantas, 1; svadhdvantas, 
voc. ; s&arvantas ; himavantas ; hr'shivantas. 

B. The forms are : aUjimantas ; ishumantas ; xshtimantas, 3 ; kshumanias, 2 ; 
jyotishmantas, 2 ; tiishimantas ; dyvmantas, 4 ; madhumantas, 12 ; midhushmantas ; 
yavamantas ; rayimantas ; vadhU'mantas (i.126.3); vd'pimantas, 2; pipwmantas ; 
sthivimdntas ; harshumantas ; havishmantas, 18. 

Weak form for strong: 'indrdvatas, iv.27.4; see p. 516. 

Nominative and Accusative Plural Neuter. 

The paradigms end in -vdnti, -m&nti. We find ghrtdvdnti, 
ix.96.13c/ papumd'nti, 97.1. «/ 92.6a. In every case, the d is 
eighth in a pdda of eleven syllables, and the pada reads -anti 
(Prat, ix.27,30). The SV. variants (i.532,526) of the first two 
passages also read -dnti. Cf. p. 510. 

Weak form for strong: kshumdti (ytithd'), iv.2.18; see p. 516. 
The AV. reads dl yHtheva kshi'im dti (!) papvd akhyat, xviii.3.23. 

Accusative Plural Masculine. 

A. There are 54 forms (from 28 stems). 

B. There are 16 forms (from 8 stems). 

A. The forms are: drratas, 13; indrdvatas (x.101.1); rtviydvatas ; farpand'- 
vatas ; jcmivatas ; tdpasvatas, 2 ; trshyd'vatas ; tvd'vatas, 2 ; nrvatas, 2 ; pdtnivatas, 
2 ; pastidvatas, 2 ; poshid'vatas ; prajd'vatas, 2 ; bhangurd'vatas, 3 ; ydpasvatas ; 
ydtiumd'vatas ; rabhasvatas ; rayivatas ; revatas ; v&nanvatas ; vrci'vatas, 2 ; par- 
yand'vatas ; pdpvatas, 6 ; pimivatas ; sutd'vatas ; s&nrtd'vatas ; hiranyavatas • 
heshasvatas. 

B. The forms are : gomatas, 8 ; jyotisfrinatas ; prashtimatas ; mddhumatas ; 
nadhU'matas (vi.21.8 : viii.57.17); vir&kmatas; st/nv/matas ; suastimdtas. For 
dyumdta i-, p. -tali, read -te, vi.lf.14. 

Strong form for weak. Possibly prdyasvanto (nd satrd'ca dl 
gata), x.lTAd, is an A.p. : 'Come hither to (us), who are united, 
(and) who have, as it were, your favorite morsels ready for you.' 
Still less probable is it that havishmanto in verse 1 is an A.p. 
Has not an effort at formal parallelism (cf. lb, 3c, 4d, 5b) dis- 
turbed the text in Ad, and displaced an original prdyasvato? 
At vi.27.8, Mtlller's first and second editions and Aufrecht's first 
have indeed vadkti 'mantas ; but Aufrecht's second reads vadhil'- 
matas, and so BR. The Atharvan, at xviii.l.57c, reads dyumd'n 
dyumantd (! so MSS. and ed.) d! vaha — a curious imitation of 
verse 56c and RV. x.16.12. The impossible accent seems to be a 
blind imitation of upatds; we must read dyumdntas or dyumdtas. 

Instrumental Plural Masculine and Neuter. 

A. Masculines: drvadbhis, 10; ddkshindvadbhis; pdtnivadbhis; 
maghdvadbhis; rdbhasvadbhis; revddbhis; vd'javadbhis; pdp- 
vadbhis; plmivadbhis; sutdsomavadbhis. Neuter : ghrtdvad- 
bhis, 2. 



522 C. R. Lanman, [Stems in 

B. Masculines: rshtimddbhis / kridumddbhis ; g6m.adbh.ig; 
barhishmadbhis / bhdnumddbhis ; vidydnmadbhis ; havishmad- 
bhis. Neuter : mddhumadbhis. 

Dative Plural Masculine and Neuter. 

A. Masculines: maghdvadbhyas, 19; from AV., bhdgavad- 
bhyas, v.31.11; mil'javadbhyas, v.22.14; sdmavadbhyas, xviii.4.73. 

B. Masculines: ydtumddbhyas (Gr., -bhias), vii.104.20,25; from 
A V., mdtrmddbhyas, xii.1.60. Neuter: vibhumddbhyas. 

Genitive Plural Masculine. 

A. Here belong : drvatdm, 4 ; ddJcshindvatdm ; bhangurd'- 
vatdm; ydtumd'vatdm, 2 ; ydpvatdm, 4 ; (Amivatdm; sutd'vatdm; 
from AV., datvdtdm, iv.3.4 ; himdvatdm, vi.95.3. 

B. The Rik has no example. The AV. has gdmatdm, iv.36.6. 

Locative Plural Masculine. 

A. Here belong : dmavatsu; drvatsu; yushmd'vatsu; maghd- 
vatsu, 1, and x.93.14 (Gr., maghdshu). 

B. Here belongs dyumdtsu. 

STEMS IN VAN, MAN, AN. 

The stems formed by the suffixes van, man, and an are declined 
alike, and it has therefore seemed best to treat them in a single 
section. The maw-stems, however, show peculiarities not shared 
at all, or only to a slight extent, by those in van and an: viz., 
the ecthlipsis of m in the I.s., the dropping of the ending in the 
L.s., and the long -a of the samhitd in the N.A.p.n. 

These stems are declined for the most part only in the masculine 
and feminine. For the few feminine forms from rc-stems and for 
the regular ways of forming the feminine, see below. The 
masculine and neuter forms from stems in van are given case by 
case under category A ; those from stems in man, under category 
B ; and those from stems in an, under category C. 

The A-stems. The stems in van are chiefly verbal adjectives 
and the forms are almost exclusively masculine. Only about a 
dozen stems show neuter forms: two adjectives, pdtvan and 
vivdsvan; and the substantives tugvan, 1 and 2 dhdnvan, pdrvan, 
a-parodn, snd'van, turvdn, ddvdn, an-arvdn, and somapdrvan. 

The B-stems. The stems in man may be pretty equally divided 
between masculines and neuters. The latter are verbal abstracts ; 
the former, nomina agentis and compounds of the neuters. 

The C-stems. The stems in an are masculines and neuters. 
They are not numerous. 

The formation of several stems is not wholly clear: e. g., 
rjipvan (A ?) ; firshdn (C), whose sh is perhaps a part of the 
suffix. Since the v of yuvan (A) is of merely phonetic origin 
(yu-v-an) the stem ought to be put under C. Under C belongs 
pu-dn; so vi-bhv-dn, pdri-jm-an. 



van, man, an.~] Noun- Inflection in the Veda. 523 

Supplementary themes. The forms of the n-stems are supple- 
mented oftener than any others by forms from auxiliary themes. 
This fact is remarkably illustrated in the N.A.s.n. of C-stems. 
Cf. Pari, vi.1.63. The cm-stems are generally oxytone ; the others, 
barytone. 

One stem appears in quadruple form : dsia, dsdn, d's, and dsa 
(a transition-stem, seen in the adverbially accented dsayd'). Two 
stems appear in triple form : dhan, dhas, dhar; Wdhan, rt'dhas, 
ti'dhar. With dsia, dsdn, d's, cf. yushd, ytishdn, yd's (TS. 
vi.3.11 1 ' 4 ); udakd, uddn, ud; updn, up (Orient und Occ. ii.242) ; 
doshdn, dds. Analogous, moreover, are the groups : akshdn, 
dkshi; asihdn, dsthi; dadhdn, dddhi; sakthdn, sdkthi; and like- 
wise pakdn, pdkrt; yakdn, ydkrt; asdn and dsrt (cf. p. 463, and 
TS. vii.4.9 1 ; the irregular t may be due to the false analogy of 
pdkrt and ydkrt). Further, we may add : dr-van, -vant; r'k-van, 
-vant; maghd-van, -vant; sahd'-van, -vant; dadhi-krd' -van, -krd'. 
The feminine ydshan is supplemented by ydshand and ydshd. 

The stems in van, man, and an distinguish the weak cases from 
the strong by lengthening the a of the suffix in the strong. 

Weak form for strong. Several stems, however, are constant 
exceptions to this rule and always keep the short vowel in the 
strong cases. They are : aryam&'n, ptish&'n, bidrdpflshd' n, somd- 
pdshd'n, rbhuksh&'n, and the fem. ydshdn. Moreover, a few other 
stems show now the short vowel and now the long. 

A. The N.d.m. yd'nd, ix.68.5, points to an exceptional weak 
form ydv&n-d, rather than to y&vdnd. Similar is the N.p.m. 
maghdnas, vi.44.12, which stands for maghdvdn-as. In x.92.14, 
Roth takes anarvd'n-am as A.s.f. with dditim. The & may be 
justified on metrical grounds, as penultimate of ajagatl In AV. 
xix.35.5, we have y'e kr'tvdno devdkrtds; but krtndvo has been 
suggested. 

B. The stem tm&'n makes tm&'n-am, -d, -e, tm&'ni and tmdln 
(but cf. tmd'nam and dtrnd'nam) ; so jemdnd. 

C. The stems ukshdn and vr'shan waver between & and d. See 
A.s., N.d., and N.p.m. From the stem tdkshdn I can cite no 
strong form with &. 

Besides the strong forms with & in the written text, the metre 
points to forms with d, where the text has d. This fact was 
touched upon by Kuhn, Beitrdge, iii.121. I have looked through 
all the 642 passages in which the forms of the A.s.m., N.A.V.d. 
m., N.p.m., and N.A.p.n. occur with written d in the penult, and 
found some 19 pddas whose cadence demands a short &. These 
are given under the proper cases : e. g. purutmd'nam, text d, 
viii.2.38&. 

Strong for weak. There is no certain instance. Possibly 
jarimd'nas stands for jarimnds; see Ab.s.m. Cf. mahdtmdnas, 
A.p.m. ; and ushd'sas, as-stems, A.p.f. 

These stems show another important peculiarity, namely, the 
syncopation of thematic 4 in the following vocalic weak cases : 
the I., D., and Ab.G s., G.L.d., A.p.m. and G.p. In the L.s. and 



524 0. R. Lanman, [Stems in 

in the N. A.d.n. this syncopation is optional in Sanskrit, and never 
occurs (with one exception — see L.s.) in the Rik, so that we may 
leave these cases out of view. 

In Sanskrit, the & of the suffix dn is syncopated in the above- 
named cases. This rule holds good also for the Rik. There are 
but two exceptions in the written text {ukshdnas and vr'shanas), 
and several disclosed by the metre (see below). The v {it) of 
vi-bhv-dn is, indeed, radical rather than suffixal ; nevertheless, the 
word is treated as a ■wan-stem so far as syncopation is concerned ; 
similarly pdri-j(a)m-a?i, etc. The combinations of consonants 
resulting from syncopation of C-forms are : kn, jn, dn, dhn, vn, 
pn, shn, sn, hn, kthn, sthn, rdhn, kshn, rshn,jjn. 

In Sanskrit, the d of the suffixes vdn and mdn, when these are 
preceded by a vowel, is syncopated. This rule holds good neither 
for the written text of the Rik, nor for the text as the metre 
shows it to have been pronounced. The diaskeuasts have some- 
times written the word without syncopation, as it was also 
spoken, e. g. mahimdnas, x.54.3 ; they have sometimes written it 
with syncopation, conformably to the later orthography, but not 
to the spoken text, e. g. sdmah kaldpe $atdydmnd pathd' , ix.86.16^ 
(for which the AV., xviii.4.60, has the orthography conformable 
to the spoken text, patdydmand) ; or they have written it with 
syncopation, conformably also to the spoken text. There are, 
then, three cases : 

Case I. The text has van, man, an, and the metre confirms it. 

Case II. The text has vn, mn, n, and the metre requires or 
admits the restoration of the syncopated vowel — v{a)n,m{a)n,{a)n. 

Case III. The text has vn, mn, n, and the metre confirms it. 

The fourth possible case, — that the text should have more than 
enough syllables, i. e. van, man, an, and the metre require vn, mn, 
n, — does not occur. 

There are, in the Rik, 65 A-forms (excluding magMnas, -os, 
-dm), 59 B-forms, and 238 C-forms in which the thematic d would 
be syncopated according to the rules of the later language. The 
sum is 362. 

Case I. In 48 instances the d appears in the written text. 
These are : (A) ddvdne, 28 ; — (B) omdnd ; pravddydmand ; 
bhllmdnd; bhU'mand; syil'mand; hemdnd; trd'mane; dd'mane; 
bhil'manas; mahimdnas, 2 ; dd'manas; bh-Olmanas, 3 ; yd'manas; 
viomanas; — (C) ukshdnas; vr'shanas (RV. iv.2.2 and AV. xi. 
2.22) ; updne?. 

Case II. In (362 — 48=) 314 instances, the a is syncopated in 
the written text. In 46 of these 314 instances, Gr. proposes to 
restore the syncopated a. 

II. a. In only 12 instances is this restoration imperative. These 




8d'm(a)nas-sdm(a)nas, ii.23.17 ; (C) mah(d)nd, text mahnd', iv. 
2.1 : x.6.7 ; vfsh(a)nas, viii.7.33. 



an.] Noun- Inflection in the Veda. 525 

II. b. The remaining 34 of the 46 restorations proposed by Gr. 
are at the end of catalectic pddas of 7 or of 11 syllables, which 
are thus converted into acatalectic pddas of 8 or of 1 2. Like the 
resolutions of -dm etc. in similar situations, they are not impera- 
tively necessary ; but they are admissible. Thus in vii.31.1c, the 
text has sdkhdyah somapd'vne; a and b are pddas of 8 syllables, 
and all three may be made uniform by reading sdkhdyah soma- 
pd'v(a)ne. Cf. Kuhn, Beitrdge, iv.195. There are 27 restora- 
tions at the end of pddas of 7 syllables: (A) drdv(a)ne, vii. 
31.5; vr'shapraydv(a)ne, viii.20.9; somapd'v(a)ne, vii.31.1 ; 32.8; 
drdv(a)nas, i.36.15 ; somapd'v(a)nas, viii.67.7 ; drdv(a)nas, A.p., 
ix.13.9;' 61.25; 63.5; rdrdv(a)ndm, viii.39.2 ; vdjadd f v(a)ndm, 
i.17.4; sahasradd'v(a)ndm,i.l f l^5; somapd'v(a)ndm,i.SO.ll; (B) 
dhd'm(a)ud, ix.39.1 : x.166.4; sd'm(a)nd, viii.84.7; usrdydm(d)ne, 
iv.32.24a/ dnusra-,24b; sd'm(a)ne, viii.4.17 ; 6.47; sushd'm(a)ne, 
viii.23.28; 24.28; 26.2; dhd'm(a)ne, viii.52.11 ; 81.25: ix.24.5 ; 
(C) rd'j(a)nd, text rd'jnd, x.97.22 ; — further, from the AV., 
dhd'm(a)nd, x.5.7-14 ; sthirddhdm(a)nas, x.4.11 ; nd'm(a)ni, 
N.d.n., iv.9.10. 

II. o. There are 7 restorations at the end of pddas of 11 sylla- 
bles: (A) svadhd'v(a)ne, vii.46.1 ; drdv(a)nas, viii.49.10; drd- 
v(a)nas, A.p., i.36.16 ; evayd'v(a)nas, ii.34.il ; (B) suhdvttund- 
m(a)ne, ix.85.65 (cf. a) ; dhd'm(a)ne, x.76.8 ; dhd'm{a)nas, i.87.6. 

Case III. In 268 instances the d is syncopated. The forms 
are : (A) grd'vnd, 2 ; -as, 2 ; -dm, 2 ; dadhikrd'vnas, 6 ; prdtar- 
yd'vnas ; bhilridd'vnas ; yuktdgrdvnas ; sutapavne; -as; su- 
dd'vne; svadhd'vne; — (B) aryamne, 2 ; -ds, 6 ; jarimn'e; nd'mnd; 
dhindmndm ; mahimnd', 3 ; -e; -ds, 2 ; Idmnas-lomnas; — (C) 
akshnds; dpnd, 3 ; -as; dhnd, 3 ; -e, 3 ; -as, 17 ; -dm, 28 ; dsnd'; 
-'e; -ds, 3 ; indrdp-Ashnds; ukshnds, 3 ; udnd! , 5 ; -ds, 6 ; ib'dhnas; 
janard'jnas; dadhnd', 3 ; -ds ; ptfshnd'; -e, 4 ; -ds, 6 ; prati- 
di'vne; mahnd', 32 ; milrdhnds, 2 ; yaknds; yamdrdjnas; ydsh- 
nds; rd'jne, 3 ; -as, 9 ; vr'shnd, 2 ; -e, 34 ; -as, 43 ; -dm, 2 ; 
pirshnd', 3 ; -'e; -ds, 3. 

In 60 instances out of 362, the d is written, or required by the 
metre. If we eqnsider only the A and the B-forms (124), we find 
that in nearly half of them (54 — from I. and II. a), the & is written 
or required, and that in the rest (70) it is syncopated. That is, 
in tabular form : 





I. 


II. a 


lift 


II. c 


III. 


Sum 


A 


28 


1 


13 


4 


19 


65 


B 


17 


8 


13 


3 


18 


59 


C 


3 


3 


1 




231 


238 



4'8 12 27 7 268 362 

Transition to the a-declension. This is common in Pali; cf. 
Kuhn, JPdli- Gr. pp. 73-5. Thus from the form addhdn-am (adh- 
vdn-am) comes the stem addhdna — I.s. addhdnena, Dhammapada, 
page 263. 

vol. x. 70 



526 C. B. Lanman, [Stems in 

A. In i.37.1, we read pdrdho (neut.) md'rutam anarvd'nam : 
we may regard the last word as a form of transition to the 
a-deelension (anarvd' na-m, A.s.n.) ; but another explanation has 
been offered on p. 330 top. Cf. also Mtiller, Translation, p. 56. 

If Both is right in taking anarva'n-am, x.92.14, with dditim (' the inviolable 
Aditi' — see p. 523), we may regard this form as the point of departure for the 
nominative anarv&'na-s, v.51.11 : viii.31.12. Here, however, Anwrvana-s seems 
to be personified as a deity, ' The Inviolable,' so that the mythology as well as 
the transition-form would indicate that these two passages belong to the upper 
strata of the Vedic deposits. The stem satvand occurs but twice (-dm, x.115.4: 
-ais, V.3T.4), and as entirely equivalent to the common satvan ; its accent, how- 
ever, makes its relation to satvan doubtful. If these are not regarded as 
transition-forms, we must assume the stems anarvdn-a, satvan-d, as secondary 
derivatives from ram-stems ; but these are very rare (nivand, vagvand, pupulcvand). 

In ii.40.6. and vii.40.4, occurs (dditir) anarvd'; this may be 
regarded, without undue violence to Vedic grammar, as a femi- 
nine nom. sing, to the stem anarvdn; and if so, it may be the 
point of departure for the neuter form anarvd-m, (i.164.2, cakrdm; 
185.3, ddtrdm), anarvd' being felt as N.s.f. to a stem anarvd!, and 
a N. A.s.n. being formed after the analogies of the a-declension. 
BR. and Gr., however, refer these forms directly to a stem anarvd. 

Six or more stems in van have subsidiary a-forms. Thus 
beside r'kvan, r'bhvan, tdJcvan, pikvan, vdkvan, vibhd'van, we 
find: ricvd-s (x.36.5), r'bhva-m (vi.49.9: x.120.6), tahvd-s (viii. 
58.13),' pikvd-s (AV. x.6.3), vdkvds (N.p.f., RV. x.148.5; A.p.f., 
iv.19.7), vibhd'va-m (i.148.1). The a-forms are infrequent. Their 
relation to the van-stems is somewhat problematic, and perhaps 
they are, after all, independent formations. Nevertheless, the 
relation of the nominatives r'bhvd and vibhd'vd to the accusatives 
r'bhvam and vibhd'vam is strikingly like that of the Fdli nomin- 
atives yuvd and muddhd to the accusatives yuvam and muddham 
(Dhammapada, verse 72). 

B. There exist beside the waw-stems a number of equivalent 
ma-stems, which are of sporadic occurrence or of later date. 
Compare darmdn with darmd-s (iii.45.2); dhdrrnan (RV.) with 
dhdrma (not till after RV.); eman with ema-s (VS. xviii.15) ; 
hdman with hdmdya (VS. viii.58); djman with djma; vi$vdkar- 
man-d etc. with vipvdkarmena (only RV. x.166.4) and Vissa- 
kammena (Dhammapada, p. 11 7 end); vrsha-karman with vird- 
karma-m ; priyd-dhdman (AV. xvii.10) with priyd-dhdmdya. 
The stem yd' man is Vedic only ; yd ma, Vedic and post- Vedic. 

C. Transition to the a-declension is most frequent with the an- 
stems, particularly in the N.A.s.n. (q.v.), where a-forms are exclu- 
sively used. In the masc, corresponding to the A.s. pUshdn-am, we 
find the transition-nominative sing. pUshdna-s (x.93.4), and the G.s. 
pUshandsya (with possibly false accent in the unclear hymn x.5, 
verse 5). For the incorrect a-pirshdnds, N.p.m., Khila to x.103, 
the AV. has -nds, vi.67.2. In other instances, the N.A.s.n. may 
serve as the point of departure ; see p. 347. Thus ftrshan makes 
its N.A.p.n. pirshd'ni or pirshd' ; and correspondingly, its L.s. 
pirshdni (RV.) or pirshe (AV. vii.56.6: xiv.1.55); its dual (due) 



van, man, an.] Noun- Inflection in the Veda. 527 

plrshe (RV.), and its N.A.s. ptrshd-m (cf. sisam). Just so dhan 
makes dhdni and dhd; and correspondingly, dhn-dm and dhdndm 
(viii.22.13). Similarly ytishdn makes ytishd'ni (TBr. Comm. 
11.668 — BR. stem ytisha) ; but in the Ls. we find y&shnd' (VS. 
xxv.9) and ytishkna (TS. v. 7.20). Starting from the weak forms 
dpn-as, ahn-ds, -i, etc., come the transition-stems dpna, ahnd 
(accent), and the forms ptirodhn'e, L.s., x.34.11 ; dpnais, viii.2.2. 
Just as the Skt. dharma usurps the place of the Vedic dhdrman, 
so vrsha (cf. vusa) supplants vr'shan. 

With especial frequency do a-stems take the place of aw-stems at 
the end of compounds: thus, akshdn, -akshd; ukshdn, mahoJcshd-m 
(Cat.Br. iii.4. I s ) ; vr'shan, -vrshd-s (AV. v.16). Compare, further, 
an-asthdn with an-asthd-s (RV. viii.1.34) and ■drv-asthd't (Cat.Br. 
viii.7.2"); brhdd-ukshd (N.s.m., RV.) with brhdd-ukshdya (VS. 
viii.8) ; rd'jan with adhirdjd-m (x.128.9); saptdbhis with tri- 
sapta'is; etc. 

Transitions from the o-declension. See nikdmabhis, Lp.m., and 
above p. 347. 

Accent. In the weak cases of oxytone stems, if the d! is synco- 
pated, the accent is thrown forward upon the case-ending. A. 
Nearly all the van-stems, are barytone, and no instance of an 
accented case-ending occurs. B. Of the wan-stems, the neuters 
are barytone ; but the masculine verbal abstracts and nomina 
agentis are oxytone, and accented case-endings are not infre- 
quent : e. g. mahimn-d', aryamn-i,, klomn-ds. C. Of the an- 
stems, some 15, masculine and neuter, are oxytone, and accented 
case-endings often occur: e. g. majjn-a", pUshn-e, mdrdhn-ds, 
indrdptishn-6s, ukshn-ds (A.p.m.); udn-d', dsn-'e, yakn-ds, etc. 

Since yiz'n-e, yU'n-as, pUn-as stand for y{ivan-e, -as, pudn-as, 
their accent is perfectly regular ; so tmdn-d, for dtmdn-d (p. 341 
end). Compare p. 408 med. 

THE FEMININE FORMS. 

A. The forms that serve as feminines to the stems in van are 
regularly made from other stems in varl. 

The RV. has 25 such stems: abhibh&'vwri ; rtd'vari; evayd'vart; talpapi'varl ; 
nishshidhvwi ; pi'vwri; p&rvajd'vart ; pras&'vart ; bdhus&'vari ; bh&ridd'vari'; 
mdtaribhvart ; yajvari ; vakvwrt ; vibhd'vwri ; pdkvaH ; pdrvart ; prushttvdri ■ 
pvetayd'vari ; sarhpipvari ; sajilvart ; sayd'vart; sumndvdri; sitnrtdvwri (also 
-vati); sr'tvart; svadhd'vwH. Further occur: agretvari, AV. xii.1.57 ; abhikr't- 
v<wt,ii.8.2; abhi'tvwri, TS. iv.1.10 2 ; nltdnapi'vwri, AV. iii.21.10 ; kr'tvari, iv.l&l ; 
mdtaripvarl, v.2.9 (cf. RV.) ; vahyapt'vari, iv.5.3; vimr'gvari, xii.l'29,35,37. For 
pretvari and Goddvari, see BR. Cf. Bollensen, Z.D.M.G. xxii.604. I know of no 
Vedic stem in vani. Exceptional is athwvi' {dfharvan). 

Only a few sporadic feminine forms are made from stems in 
van, and these I regard as irregular extensions of the wa«-stems 
beyond their proper field (p. 515-6), due perhaps to the false 
analogy of the maw-stems. These forms are : sayugvd (gdyatrf), 
x.130.4; anarvd' (? dditis), ii.40.6: vii.40.4; anarv&'n-am (? ddi- 



528 C. R. Lanman, [van, man, cm-stems. 

tim), x.92.14; sa(m)sthd'vdnd (rddasi), viii.37.4; indhanvabhis 
(dhewdbhis), ii.34.5; sahasradd'v(a)ndm (sumatind'm), or m., 
i.17.5. 

B. So far as the evidence goes, it seems to show, on the other 
hand, that stems in man originally served for all three genders. 
Forms like sutdrmdnam {nd'vam), therefore, I regard as regular. 
The stems in man seldom offer occasion for the formation of fem- 
inines, save at the end of compounds. 

There is not a single certain example in the Eik of a feminine formed by adding 
i to the stem in man; but the AV. has several: durnd'mni, iv.17.5 : xix.36.6; 
pdficandmni, viii.9.15 ; Tnahd'-, xi.7.6 ; vipud-, ■m.16.2 ; sahdsra-, viii.7.8. Gr. 
refers tmdnid to tmdni, fern, of tmdn; of. BR. Doubtful is dwadmami' (of. 
adman), VS. ii.20. The entirely exceptional fem. to brahman, namely brdhmt, is 
found only in a play upon words with yahvl' (BR.), ix.33.5. 

The feminine forms from man-stems are: sdlakshmd, x.10.2; 
12.6; dyutddydmdnam (Kshdsam), v.80.1; sutdrmdnam (nd'vam), 
viii.42.3 ; sutrd'mdnam (prthivi'm), sutdrmdnam (dditim), x.63.10 
(=AV. vii.6.3) ; drishtabharman (adite), viii.18.4; prthuydman 
(duhitar), vi.64.4 ; sujdnmani (dhishdne), with neuter ending ! — 
cf. p. 433 top, i.160.1 ; pucijanmanas (ushdsas), vi.39.3 ; vd'ja- 
bharmabhis (utibhis), viii.19.30; sukdrmabhis (Say. angulibhis), 
ix.70.4 ; pukrdsadmandm (ushdsdm), vi.47.5 ; from AV., su- 
sh-Olmd, vii.46.2 (Gr. and BR., stem -ma). 

C. The word ydshanas, 'women,' occurs 8 times; aside from 
this there are 6 feminine forms from am-stems. These are : vr'shd, 
(kd$d) viii.33.11 : (vd'k) x.115.8; vr'sh&nam (tvdeam), i.129.3; 
vr'shand (dyd'vdprthivi'),x.66.6; pdrijmdnas (vidydtas), v.10.5 ; 
rappadttdhabhis (dheniibhis), ii.34.5. 

The aw-stems appear at the end of feminine compounds with the suffix i as 
follows: in the Rik, achidra-Adhni ; samr&'jni; somardjni; hatavrshni; r&ru- 
ftrshni; sapt&firshni ; in the AV., eka/mS/rdhni, viii.9.15; sindhurdjnt, vi.24.3. 

THE MASCULINES AND NEUTERS. 

The feminines being thus enumerated, we may proceed to the 
masculine and neuter forms. 

Nominative Singular Masculine. 

A. There are 308 forms in -vd (from IS stems). B. There 
are 207 forms in -md (from 49 stems). C. There are 376 forms 
in -d (from 18 stems). Sum, 891. 

According to Curtius, Studien, ii.163, dpnd and pitd' presuppose the forms 
dprndn and pitd'.r. Continuing, then, the search begun on p. 424 top, I examined 
every one of these 891 forms as it stands in the samhitd and obtained the results 
here given. Collision of final -d with a vowel occurs 99 times in the interior of 
a pdda, and about 25 times at the end of a pdda (e. g. iii.1.23 : iii.5.9 : vi.22.7 : 
vi.12.4). 

In the interior of a pdda, the vowels are written as coalescing in 74 instances, 
and the metre also shows that they must be combined. Thus -d unites with a- 
28 times : as in i.32.3; with i- (especially the i of iva) to e 42 times : as in ii.26.1 ; 
with u- to o in i.161.8 : iii.27.14 : v.46.5 ; with e- to ai in x.61.26. 



tN.s.m.] Noun- Inflection in the Veda. 529 

The vowels are written as coalescing, where the metre demands hiatus, in 18 
instances: i.69.96; 87.4c; 91.56; 174.1a; 178.2a: ii.20.3a?: iv.40.2a"; 57.76: 
v.31.12c; 44.3d: vi.20.36 : vii.40.46 ; 93.7a" : viii.16.7a: ix.86.45a ; 88.3d!: 
x.6.26; 117.7c. 

In only four passages have the diaskeuasts taken due account 
of the metrical value of -d a- as two syllables and written them 
with hiatus: namely, in the combination pilshd' a-, v.51.11 : 
vi.50.5: x.26.1,9, covered by the Prat., ii.28. There are only 
three instances in which r- follows. They are 

iv.33.3. te vd'jo mbhvdfl rbhur ind/ravantah, p. -vd 
iv.36.6. yam vd'jo vibhvdfo rbhdvo yarn d'vishuh, p. -vd 
vii.48.3. indro vQihvM rbhukshd' vd'jo arydh, p. vd. 

The precept of Qaunaka (H.31) excepts vibhvd from the influence 
of ii.ll, whereby we should have vibhvd r-. According to the 
metrical canon (p. 423-4), the combination -d r- results regularly 
in or, except . ..." (3) when . . . -d stands for . . -ds, -e, -ai, 
-dn, or -dr." Since the metre here demands hiatus or at least 
forbids the fusion vibhvarbhtir etc., Grassmann ( Wb., p. vii) 
would " restore the original form vibhvdn." To these may 
perhaps be added r'bhvdn, vi.34.2, text r'bhvdn at the end of 
a pdda (before ekah) ; see Prat. ii. 31, r. 61, clxvi. 

A. The forms are: agrayd'vd; dtharvd, 5; admasddvd; ddhvd, 8; anarvd', 
7; abhifastipd'vd, 2 ; abhisatvd; abhiardhaydjvd ; amativd' ; ardtivd', 4 ; drdvd, 
5; drvd, 20 times, and vii.58.4; dfupatvd; ugrddhanvd; r'kvd, 2; r'ghdvd ; 
rjifvd, 2; rnayd'vd; rndvd' ; rtd'vd, 31; r'bhvd,1; krshndadhvd, 2; kratwprd'vd; 
kshiprddhanvd ; grd'vd, 16 ; tdkvd ; dadhikrd'vd, 3 ; devayd'vd ; drushddvd ; 
nrshddvd; pdtharvd; pdtvd: purukr'tvd; p&rvagdtvd; ptirvayd'vd ; prarikvd; 
prdtaritvd ; bddhasr'tvd ; bh&ridd'vd ; maghdvd, 68 ; mddvd ; mrakshakr'tvd ; 
ydjvd, 3; yd'vd; yuktdgrdvd, 2; ytidhvd, 3; yiivd, 30; raghupdtvd; rdrdvd; 
vdkvd, 2; vasudd'vd; vdjadd'vd; vijd'vd; viWid'vd, 19; vibhr'tid; vr'shaparvd ; 
pubhamyd'vd ; ptibhvd ; (yendpatvd ; prutar[a]vd ?; frushtivd' ; satindsatvd ; satyd- 
madvd; sdtvd, 10 times, and v.33.5 ; sarnadvd, 2 ; sahd'vd, 4; s&tvd; sudhdnvd; 
sr'tvd; somapd'vd; somasuivd; stitbhvd; hilvdf. 

B. The forms are: akarmd' ; dbrahmd, 2; aryamd', 77 ; apmd, 4; asremd' ; 
dtmd', 15; dfuhemd, 4; ishtdydmd; usrdydmd; kridbrahmd, 3; jarimd', 6; 
jdt&'bharmd ; trpdlaprabharmd ; darmd' ; ddmd' ; dwnd'md, 2; dwmdnmd; 
dyutddydmd; dvijdnmd, 3 ; dharmd' ; dhvasmd' '; pd'kasihdmd, 2; prtMpragdmd ; 
brahmd', 24; bhujmd'; bh&'rijanmd ; mahimd', 16; yajndmanmd ; raghuyd'md; 
vadmd',2; varimd'; vipodkarmd, 3 ; vr'shaprabharmd ; fatd'tmd,2; satyddharmd, 
3; satydmanmd, 2; saptandmd; sahdsraydmd; sddMikarrad ; sukdrmd; mjdnvmd, 
3 ; sutrd'md, 4 ; sudyotmd ; svhrahmd ; swmdntvmdmd ; sumdnmd ; suvdhmd ; 
sushthd'md ; svddukshddmd. 

C. The forms are: anasthd' ; dyahftrshd; afirshd' ; ukshd', 9; takshd; tdpw- 
mtirdhd, 2; tritidhd' ; dvibdrhajmd; pdrijmd, 13; pilshd', 56; brhadukshd; 
mdtarifvd, 19; miurdhd', 8; rd'jd, 99; vibhvd, 10; vr'shd, 150: fvd', 2; sahdsra- 
firshd; from AV., majjd', iv.12.3,4; from YS.,plihd', xix.86. 

Besides the regular nominatives arvd, maghdvd, and sahd'vd, 
there are subsidiary forms in -ban; see p. 516. In some cases 
drdvd is only a metrically suitable way of writing drvd (cf. hdri- 
pmapdru, x.96.8, with pmdpru) ; so in vii.68.7 ; and perhaps in 
x.40.7, yuvdr drdvd, text yuv6 rdrdvd. In ix.21.5, c is catalectic 
(not drddvd). In viii.63.4, read prut-dr[a\vd ? In AV. iv.35.5, 



530 0. B. Lanman, [van, man, aw-stems. 

the reading prdnadd'vdn seems to require a correction to -vd. 
For bhtijmd, RV. i.65.5, BR. propose bhujrnd'. For krdnd', see 
pp. 334, 329. 

The word ghrdns occurs AV. vii.18.2, nd ghrdns tatdpa. BR. 
set up a stem ghrdns. Whitney, second marginal note to Ath. 
Pr. ii.26, regards the s as inserted. Is not the stem formed with 
suffix an (ghdr-an, ghr-dn) ? Its monosyllabism after syncopation 
might then account for the anomalous case-form. 

Transitions to the a-declension. For anarodna-s, pilshdna-s, 
etc., see p. 526. 

Nominative and Accusative Singular Neuter. 

The pada-iorm. of the case always ends in -& ; but the final is 
sometimes lengthened in the sarhhitd, chiefly for the sake of the 
metre. See Whitney's note to Ath. Pr. iii.16. I shall give under 
I. the forms having -d in both texts ; and under II., such as are 
lengthened in the sarhhitd. 

L A. Here belong : dhdnva; dhdnva, 6 ; vivdsva; from AV., 
pdrva, i.12.2 ; snd'va, xi.8.11,12. 

B. There are 420 forms (from 35 stems). 

The forms are : ajma, 4 ; ddma ; ema, 3 ; karma, 5 ; kd'rshma ; kshadma, 2 
carma, 6; jdnima, 4; janma, 9; tokma; dd'ma, 2; dhdrma, 3; dhd'ma, 31 
nd'ma, 78 ; pdtma; brdhma, 78; brdhma-brahma ; bhU'ma, 11; manma, 33 
marma, 3 ; yd'ma, 2 ; v&rma, 11 ; vdsma ; v&gma, 2 ; vioma, 2 ; pAkma ; forma, 
84; sadrna, 14; sd'ma, 10; sd'ma, 3; sthd'ma; sy&'ma; svd'dma, 2 ; homo. For 
U, see N.s.m. ; for vr'shand'ma, see below. 



C. My collections yield the interesting negative result that not 
a single neuter stem made by the derivative suffix an forms a 

N.As. 

If we derive kshd'man from ksliam (ksM'm-an), we shall have to admit one 
exception, for kshd'ma occurs five times; but it may come from kshd=kshi 
(ksM'-man), and at any rate'the word was felt as a man-stem — not as an are-stem. 

It is no mere accident that these forms do not occur, since the occasion for 
them is not infrequent. It may be that a form like firsha' was to the apprehen- 
sion of the language-users too much like a naked stem and was accordingly 
avoided. But whether we can motivate this avoidance or not, there is no 
question about the fact that they were avoided. The forms from stems in man, 
on the other hand, are frequent and familiar. 

The N.A.s.n. in -a from are-stems is avoided by recourse to 
other stems. The stems dhan and ii'dhan have as a N.As.n. 
dhar and H'dhar ; akshdn, asthdn, dadhdn, and sakthdn have 
dkshi, dsthi (AV. iv.10.7), dddhi (TS. ii.5.3 4 ), and sdkthi (TS. 
v.3.12 a ) ; dsdn, uddn, pirshdn, and ytishdn have dsia-m, udakd-m, 
piras (and ptrshd-m, AV. iv.34.1), and yHsha-s (Schol. to VS. 
xxv.9) ; yakdn, pakdn, and asdn have ydkrt (AV. x.9.16), pdkrt, 
and dsrt (TS. vii.4.9 l ). The last form (for dsrk) has been regarded 
as a purely phonetic anomaly (p. 466) ; but see p. 523. Instead 
of y&vd from ydv-an we find ydvat as if from yd-vant; see p. 516. 
Furthermore, the form vr'shd as a neuter adjective is avoided in 
several ways : either by putting the masculine form with a neuter 



N.A.s.n.] Noun- Inflection in the Veda. 531 

noun (vr'shd vdnam, ix.64.2 — cf. gotrdm haripriyam, p. 3*1*1) ; or 
by using the derivative vr'shniam (with pdvas, viii.3.8 : ix.64.2: 
Val. 3.10) ; or by using the otherwise unheard-of vr'shni (paroxy- 
tone — with cdvas, v.35.4 : viii.3.10; with paiinsiam, viii.^S). 
For vr'shand'ma (ix.97.54 — corrupt), Gr. suggests vr'shano nd; 
see Ueb. 

These surrogate-forms are to be regarded, for the most part, as 
supplementary rather than transitional. But see p. 526. 

II. The samhitd has -d; pada, -d. Here belong : (A) dhdnud; 
(B) kdrmd; jdnimd; bfv&'md, 3 ; viomd; svd'dmd; dhd'md, 
AV. vi.31.3 ; (C ?) kshd'md, 3, and AV. xviii.3.21 (RV. iv.2.16). 

This protraction occurs in the second place of a trishtubh-pdda in iv.2.16: 
x.45.4; of an anushtubh-pdda, X.1T6.1; of an aksharapankti, i.69.3d; in the 
eighth place of a trishtubh-pdda, i.62.8 : x.129.1 ; of a jagati-pdda, ii.24.14: 
x. 142.2 ; and as penultimate of a trishtubh-pdda, vi.12.5. In all these eases it is 
metrical. It occurs also in the fifth place of a trishtubh-pdda, i.61.14 (Prfit. 
viii.8) ; and at the end of a trishtubh-pdda, i.173.6 (viii.30). The AV., at vi.31.3, 
has -d in the fourth place, while the RV. (x.189.3) has -&. 

Transitions to the a-declension. For anarvd'na-m (?), anar- 
vd-m, ptrshd-m, etc., see above and p. 526. 

Accusative Singular Masculine. 

I. The A.s.m. generally has the penult long. A. There are 
68 forms in -vdnam (from 23 stems). B. There are 85 forms in 
-mdnam (from 32 stems). C. There are 51 forms in -dnam 
(from 12 stems). Sum, 204. 

A. The forms are : adhvdnam, 3 ; anarvd'nam, 1 ; ayqjvdnam ; drvdnam ; 
Xjipidnam ; rndvd'nam ; rtd'vdnam, 5 ; grd'vdnam, 2 ; dadhihrd'vdnam ; dhitd'- 
vdnam, 2 ; pt'vdnam ; purunishshidhvdnam ; puroyd'vdnam, 3 ; prdtaryd'vdnam, 
3; maghavdnam, 11; mushivd'nam; yuvdnam, 16; frushtivd'nam, 2 ; sajitvdnam; 
satvdnam, 2 ; sayd'vdnam ; sahd' vdnam ; sr'lcvdnam. 

B. The forms are : dpmdnam, 8 ; asremd'nam ; dtmd'nam, 4 ; ojmd'nam ; 
omd'nam, 4 ; jarimd'niam ; tuvibrahmdnam ; darmd'nam, 2 ; ddmd'nam, 3 ; dur- 
mdnmdnam; dyutddydmdnam ; dvijanmdnam; dharmd'nam, 2 ; pd'kasthdmdnam ; 
purutmd'nam ; brahmd'nam, 5 ; bh&md'nam ; mahimd'nam, 21 ; varimd'nam, 2 ; 
varshmd'nam, 3 ; vifvakarmdnam ; visarmd'nam ; fatd'tmdnam ; satyddhar- 
mdnam ; sadmd'nam, 2 ; sahasdmdnam ; sudyotmdnam ; subrahmdnam ; suf&r- 
mdnam ; somd'nam ; svddmd'nam ; harimd'nam, 3 ; from AV., purundmdnam, 
vi.99.1 ; stdmd'nam, v.13.5. 

C. The forms are: ukshd'nam; trimurdhd'nam ; trifirshd'nom, 2 ; parijmdnam, 
4; prfhugmd'nam ; majj&'nam. ; mdtarifvdnam, 2 ; murdhd'nam, 10 ; rd'jdnam, 
22; vr'shdnam, 2; fvd'nam, 3; saptdftrshdnam, 2 ; from AV., plihd'nam, iii.25.3 ; 
pdnthdnam, xii.1.47, see p. 441. 

The scansion of the passages in which these 204 forms occur 
shows that the metre demands or favors a short & as penultimate 
in the cadence of ten octosyllabic pddas : in ii.6.56 and viii.81.8a 
(anarvd'nam?); iii.27.2c and 40.3a (dhita vdnam?); viii.54.4a 
(mahimd'nam ?) ; i.50.12a (harimd'nam?); viii.61.10J (parij- 
mdnam?); ix.114.2c (rd'jdnam ?) ; and possibly in Tiii.2.385 
(purutmd'nam?) and 46.3a (mahimd'nam?). In no other cases 
does the metre give evidence of a short vowel. 



532 C. H. Lanman, [van, man, cm-stems. 

I. a. From the stems that waver between & and d in the strong 
cases occur: ukshd'nam, i.164.43 ; vr'shdnam, ix.34.3 : x.89.9 : 
VS. xx.40 (in all these passages, the d being in the second or 
fourth place may be ascribed to the metre) ; ukshd'nam, VS. 
xxviii.32 : Ait. Br. i.15 ; vr'shdnam, Cat.Br. i.2.5 1B ; tmd'nam, 
Acv. Or. vi.9 ; tdkshdnam (cf. Schol. to Pan. vi.4.9), Kath. xii.10, 
in Ind. Stud, iii.464. 

II. The text has & in the penult. There are 98 forms (from 6 
stems). In no case does the metre require d; but it favors d in 
the second place of the verse in vi.55.4«: i. 16.15: ii.l6.5c#: iii. 
27.15a: viii.15.4c>: ix.63.21«; 106.15. 

The forms are : aryamd'nam, 15, and AV. xiv.1.17 ; p-dshd'nam, 
21, and AV. xi.6.3 : xviii.2.53 ; rbhukshd'nam, 4 ; further (cf. 
I. a), vr'shdnam, 53 ; ukshd'nam, 4, and AV. iii. 11.8 ; tmd'nam. 
In Dhammapada, 355c?, we have d as penultimate, hanti anne va 
attdnam (dtmd'nam). For pdnthdnam (? i.l2V.6^A : v.10.1 : 
viii.57.13), see p. 441. For anarvd'nam, see p. 531 end, and p. 
527 end. Cf. vrtrahd'nam. 

The form mahd'm is explained by Brugman, as standing for 
mahdn-M. ; see Studien, ix.308. 

Transition to the a-declension. For adhi-rd'ja-m etc., see p. 527. 

Instrumental Singular Masculine and Neuter (see pp. 524-5). 

A. Masculines : dtharvand; abhiyugvand; rjipvand, 3 ; cikit- 
vdnd; patdparvand, 4 ; — (III.) grd'vnd, 2. 

Neuters : dhdnvand, 3 ; dhdnvand; pdrvand-parvand; — (III.) 
snd'vnd, AV. vii.50.9. 

B. Masculines: demand, 2; tmdnd, 61; bhdsmand, 2; maj- 
mdnd, 22 ; varshmdnd; vipvdkarmand; — (I.) omdnd; pravdd- 
ydmand; bhtimdnd; — (II. a) patdydm(a)nd; the AV., xviii.4.60, 
and the SV, i.557 : ii.502, actually have patdydmand; — (III.) 
mahimnd', 3 ; from AV., aryamnd', ii.36.2 : xiv.1.34 ; pdmnd', 
v.22.12; varimnd', iv.6.2 : ix.2.20 ; mahimnd', iv.35.3 : vi.71.3 : 
xi.1.19: xiii.1.8*: iv.30.8 (=RV. x.125.8, mahind'). 

Once varimdn makes varimdtd; see p. 516. 

Neuters : kdrmand, 8 ; jdnmand, 5 ; dhdrmand, 15 ; pdtmand; 
brdhmand, 31 ; mdnmand, 8 ; vdrmand; vidmdnd, 4 ; vidhar- 
mand ; pdkrnand, 3; pdrmand, 5; pd'kmand; svdjanmand ; 
hdnmand, 5 ; — (I.) bhU'mand; syU'mand; hemdnd; — (II. a) 
nd'm(a)nd ; (II. b) dhd'm(a)nd, 2 ; sd'm(a)nd ; from AV., 
dhd'm(a)nd, x.5.7-14 ;— (III.) nd'mnd; from AV., dd'mnd, vi. 
103.2,3 ; dhd'mnd, v.27.8 ; Idmnd, iv.12.5 ; sd'mnd, thrice. 

C. Masculines: mdtaripvand, ix.67.31 ; AV. v.10.8; vibhvdnd 
(-dnas, Ab., needed), BV. x.76.5 ;— (IT. b) rd'j(a)nd;— (III.) dpnd, 
3; pdshnd'; vr'shnd, 2; majjfldl , AV. iv.12.3,4; pUhnd ', VS. 
xxv.8; tdkshnd, Katy. Cr. vi.1.5; — finally, fund, for $udnd, AV. 
vii.5.5. 

Neuters : (II. a) mah(d)nd, 2 ; — (III.) dhnd, 2 ; dhnd-ahnd; 
dsnd'; udnd', 5; dadhnd', 3; mahnd', 32; pirshnd', 3; from VS., 
$aknd', xxxvii.9 ; sakthnd', xxiii.29. 



I.s.m.n.] Noun-Inflection in the Veda. 533 

Shortened forms of the instrumental singular. These occur 
only with stems in m&n. After syncopation of the vowel &, (1) 
the m is dropped ; (2) but sometimes the n is dropped. 

1. The m is dropped. The complete history of the form is illus- 
trated by the stem bhiimdn; this makes: bhtl-mdn-d, RV. i. 110.2 ; 
bhU-mn-d', 9at.Br. ii.3.4 9 ; bhU-n-d', RV. In most cases the ecthlip- 
sis of m may be ascribed to the metre: thus in iii.59.7a (abhi yd 
mahind' divam) : viii.12.23 ; 57.3; 81.23: x.119.8, the dropping 
of m gives the needed diiambic cadence; similarly iii.30.13c, etc. 
See Benfey, Abh. d. kbn. Ges. d. Wiss. zu Gdttingen, xix.233 ffi 

The forms are: mahind' (cf. mahimnd', thrice), 33 times, and 
iii.7.10: viii.59.6 (=SV. ii.213); 81.23; cf. VS. xxvii.26=TS. 
iv.l.8 e : VS. xvii.l8=TS. iv.6.2': VS. vii.19: TS. iv.3.13 6 : SV. 
ii.1011,1123;— bhtind', x.149.3 ; 82.4=TS. iv.6.2 a =VS. xvii.28 
(Mahtdh., bhUmnd) ;—prathind' , i.8.S = SV. i.166; Val. 8.1;— 
prend' (cf. premnd, MBh.), x.71.1: TS. v.5.2 1 ; varind! , TS. i.5.3' 
(cf. VS. iii.5) : iv.l.3'(=VS. xi.29). It is interesting to observe 
that the VS. in the last two passages has the more grammatical 
varimnd'. In like manner the Atharvan (at iv.30.8) corrects the 
antiquated mahind' of the Rik (x.125.8) into mahimnd', to the 
detriment of the metre. Benfey would restore mahind', AV. 
iv.30.8 and RV. i.59.7. Perhaps mahnd' (34) is a shortened form 
for mahind'. Here belongs, according to Gr., dd-n-d' for dd- 
mn-d', v.52.14,15; 87.2; viii.20.14 (but see p. 335 ad init.) : 
viii.33.8 (see p. 330 end). Of. G.p.m. (mahi'ndm). 

2. The n is dropped. I have but two examples : drdgh-m-d' 
for drdgh-mdn-d, x.70.4 (Say., drdghimnd) ; rap-m-d', for rap- 
mdn-d, vi.67.1 (Say., rapmind). 

Transition-forms. For y-dsmna, etc., see p. 527. 

Dative Singular Masculine and Neutee (see pp. 524-5). 

A. Masculines: ddhvane ; dpapcdddaghvane ; rjipvane, 2, 
kr'tvane, 2 ; jdsvane; drdhvane, 2 ; prshthaydjvane; mddvane; 
ydjvane; prutdrvane; sdtvane, 2; sukr'tvane'; sthirddhanvane; 
— (II. a) rtd'v(a)ne; (II. b) drdv(a)ne; vr'shapraydv(a)ne; soma- 
pd'v(a)ne, 2; (it. c) svadhd'v(a)ne;— (III.) sutapd'vne; sudd'vne; 
svadhd'vne; — finally, yd'ne. 6. 

Neuters : turodne, 5 ; — (I.) ddvdne, 28. 

B. Masculines : tmdne, 6 ; brahmdne, 2 ; bh-d'rikarmane; from 
AV., takmdne, 1 ; pdpmdne; — (II. a) aryam(d)ne; (II. b) itsrdyd- 
m(a)ne; dnusra-; sd'?n(a)ne, 2; sushd'm(a)ne,' 3 ; (II. c) mhdvt- 
tundm(a)ne; — (III.) aryamnk, 2 ; jarimn'e; mahimnk 

Neuters : kdrmane; jdnmane, 1 ; dhdrmane, 3 ; brdhmane, 9 ; 
bhdrmane; vidmdne, 2 ; vidharmane; pdrmane; — (I.) trd'mane; 
dd'mane; — (II. a) dhd'm(a)ne; (II. b) dhd'm(a)ne, 3; (II.' c) 
dhd'm(a)ne. 

C. Masculines: pdrijmane; vibhvdne; mdtaripvane, 3; — (I.) 
np-dn-e, vi.20.ll, according to Benfey, Orient und Occ. ii.242; — 

vol. x. 71 



534 C. R. Lanman, [van, man, cm-stems. 

(III.) pttshnk, 4; pratidi'vne; rd'jne, 3; vr'shne, 34; — finally, 
pane, AV. i.11.4 : vi.37.3 : xi.2.2. In RV. iv.3.6, Bollensen reads 
ickshne, text ksh'e. 

Neuters: dhd'rv-ane; — (III.) dhne, 3; dsnb; pirshn'e-ptrshne. 

Transition-forms. For brhddukshdya, etc., see p. 527. 

Ablative Singular Masculine and Neuter (see pp. 524-5). 

A. Masculines : pdkasMvanas; — (II. c) drdv(a)nas. 
Neuters : pdrvanas; dhdnvanas, AV., 4 times. 

B. Masculines : dpmanas, 4 ; dtmdnas, 2 ; satyddharmanas, 
AV. i.10.3 ; — (III.) from AV., klomnds, ii.33.3 : ix.8.12 ; jarimnds, 
xviii.3.62 ; varimnds, xii.5.72. 

Neuters : cdrmanas, 4; jdnmanas; dhdrmanas, 3; mdnmanas; 
vdsmanas; sddmanas; — (I.) bh&'manas; — (II. a) dd"m(a)nas; 
— (III.) Idmnas-lomnas; from AV., dd'mnas-ddmnas, vii.83.2. 

0. Masculines : Benfey regards vibhvdnd c-, x.76.5, as an Ab. 
(-d=-ah) ; see Gott. Abh. xix.261 ; — (HI.) nvdrdhnds; vr'sknas; 
plihnds, AV. ii.33.3. 

Neuters : (III.) akshnds; dhnas; dsnds, 3 ; udnds; yahnds; 
ptrshnds, 2 ; vr'shnas tldhnas, iv.22.6. 

In x.21.21, it is hard to make jarimd'nas the subject of taranti. Roth suggests 
that this may be a metrical way of writing jarimnds ; cf. ardvd (=arvd) ; hari- 
prnafdru (= -pmapnt) ; sthd'tdrd (I.s.m. ? x.59.1). ' There, beyond (the sun), is 
another path ; over it (tdd), with steady tread, away from old age, men pass ; ' 
i. e., there they lay aside decay and infirmity. The construction is hardly Tedic ; 
but cf. the Sanskrit vayam Mrnd mahdbhaydt, Hariv. 4066. See also p. 523. 

Supplementary forms : dsidt, ds-ds; udakd't; etc. 

Genitive Singular Masculine and Neuter (see pp. 524-5). 

A. Masculines : dtharvanas, 2 ; ddhvanas, 6 ; dyajvanas; 
dstrtayajvanas ; durgr'bhipvanas ; ydjvanas, 3 ; viadhvanas ; 
maghdnas, 8; in v.27.1, Delbrtlck reads -dm; — (I.) magJidnas, 
pronounce maghd-van-as, v. 16.3: ix.32.1; for vi.44.12, see N.p.m. ; 
— (II. b) drdv(a)nas; somapd'i)(a)nas; — (III.) dadhikrd'vnas, 6 ; 
bhtiridd 'vnas ; yuktdgrdvnas ; sutapd'vnas, viii.2.7, troch. ; — 
y-Qlnas, 2. 

Neuter : dhdnvanas. 

B. Masculines: amdrmanas, 3; brahmdnas ; vipatmanas; 
vipramanmanas; pUcijanmanas; supdrrnanas; — (I.) mahimdnas, 
2; — (II. b) from A V., sthirddhdm(a)nas,xA.ll; — (III.) aryamnds, 
6 ; mahimnds, 2; jarimnds, AV. vii.53.5. 

Neuters: kdrmanas, 2 ; jdnmanas, 3 ; dhdrmanas; brahmanas, 
with voc, 18; brahmanas, 4; with -pdti, 32; mdnmanas, 2; 
vdrmanas; — (I.) dd'rnanas; bhiX 'manas, 3 ; yd'manas; viomanas; 
— (II. a) dhd'm{a)nas; sd'm(a)nas; sd'm(a)nas-sdm(a)nas; (II. c) 
dhd'm(a)nas;— (III.) sd'mnas, AV. xi.7.5. 

C. Masculines : pdrijmanas; — (I.) vr'shano nd, ix.97.54, text 
vr'shand'ma — see ITeb. ii.513; from AV., vr'shanas, xi.2.22 ; — 
(III.) dpnas; ukshnds; pUshnds, 6 ; yHshnds, or n. ; rdjfias, 8 ; 
vr'shnas, 37; tdkshnas, <JJat.Br. iii.6.4 4 ; — finally, p-dnas, 2. 



Gr.s.m.n.] Noun- Inflection in the Veda. 585 

Neuters : (III.) dhnas, 16 ; udnds, 5 ; dadhnds; vr'shnas, 2 ; 
pirshnds; from AV., asnds, v.19.3; asthnds, iv.12.1; paknds, 
xii.4.4. 

Transition-form: pijbshandsya ? ', see p. 526. 

Locative Singular Masculine and Neuter. 

In the Rik there is but a single instance of syncopation of the 
short thematic a in the L.s. : this is indrdgni pataddvni, v.27.6a; 
and even here we may restore the a, pata-dd'-v(a)n-i; so in AV. 
iv.7.5c, tishthd vrlssheva sthd'm(a)ni. 

From the Atharvan, I have noted a few instances : vijd'mni, 
vii.76.2 ; ISmni-lomni, ii.33.7; sthd'mni, vi.77.1 : vii.96.1; dhni, 
vi. 110.3: xix.56.2. 

All the locatives singular from the Rik may be put under two 
categories: (1) those with the case-ending i; (2) those which 
have dropped the case-ending. The former number 127 (from 48 
stems) ; the latter, 203 (from 45 stems) : that is, the forms without 
the ending are to those with the ending in the ratio of 8 : 5. 

This dropping of the ending is peculiarly Vedic. For a similar 
thing in the G&thas, see Yap. xxxi.13, cashm'eng, 'in (thine) eye.' 
I examined the passages in which the above 330 forms occur, and 
found that the choice between the two forms was often decided 
simply by the metre (cf. pp. 340,345,347). Thus in i. 143.2 and 
vi.8.2, pddas b, c, and d are each of 12 syllables, and accordingly 
we have in a, sd jd'yamdnah parame viomatii; but in vii.5.7, 
since b, c, and d are each of 11 syllables, we have in a the shorter 
form, sd jd'yamdnah parame moman. Compare also ii.28.3a 
with i.4.6c. A comparison of RV. x.87.17 with AV. viii.3.17 is 
instructive. The Rik has tdm pratydncam arcishd vidhya mdr- 
man; this the Atharvan converts into an acatalectic pdda by 
putting mdrmani in place of the antiquated mdrman. 

At the end of a pdda, the <m«-form is used if the, number of 
syllables in the other pddas is even ; and the arc-form, if it is odd. 
In the latter case, the ending can always be restored without 
detriment to the cadence, the only difference being that the verse 
thus becomes acatalectic instead of catalectic. The form in -ani 
is never found with the postposition d'; but the form in -arm, p. 
-an, occurs with it 17 times. See Kuhn's remarks upon the forms 
in -ann, Beitrdge, iii.125. 

Of the 127 forms in -ani, the great majority, 92, stand at the 
end of pddas of 12 or 8 syllables, and the other 35 in the interior 
of the pdda. Of these 92, I found 46 in Ja^a^stanzas ; one 
(i.164.366), in a stanza of 11,12,12,12; and the rest in stanzas 
consisting of octosyllabic, or of dodecasyllabic pddas, or of both, — 
variously combined (e.g. i.130.3; cf. the occurrences of pdrmani). 

Of the 203 forms in -an, nearly one-third, 62, stand at the end 
of a pdda of 11 syllables, and two (viii.2.8 : x.105.9) at the end 
of a pdda of 7. In 60 instances of the 62, the other pddas of 
the re are also trishtubh. 



536 C. B. Lawman, [van, man, cm-stems. 

In i.164.410" (12,12,12,11), we might restore the longer gram- 
matical form vioman{i), and so perhaps in il 64.39a (11,11,12,12). 
In like manner, x.105.96 admits the longer form yajndsya dkCurshh 
sddman(i); and in viii.2.8, Gr. would read samdne ddhi bhd'r- 
man(i). Unless we regard x. 6. la as a syncopated verse, we may 
read : aydrh sd ydsya pdrman(i) dvobhis. 

1. The locatives in -ani (127, from 48 stems). 

A. Masculines : dtharvani ; ddhvani, 2 ; rjipvani ; rtd'vani ; prutdrvani ; 
sukr'tvani. 

Neuters : tugvani ; pdrvani-parvani. 

B. Masculines : dpmani, 3 ; dtmdni ; tmdni, 2 ; brahmdni ; sushd'mani ; from 
AV., mahmdni, x.2.6. 

Neuters : kdrmani-karmani ; janmani ; dd'mani ; dtrghaprasadmani, 2 ; dhart- 
mani, 2 ; dhdrmani, 4 ; dhd'mani ; pdrtmani ; prdbharmani, 2 ; prdydmani ; 
brdhmani ; mdnmani ; mdrmani ; yd'mani, 12 ; yd'mani-ydmani ; vdkmani ; 
vidharmani, 8 ; viomani, 8 ; pdrmani, 11 ; sdrtmani ; sdvimani, 5 ; sushdmdni ; 
stdrtmani ; hdvtmani, 3; homani ; homani; from AV., djmani, xi.10.22 ; cdr- 
mani, vepmani, etc. 

0. Masculines : mdtaricvani ; mArdhdni, 6 ; rd'jani, 4. 

Neuters : dhani, 4 ; dsdni, 1 ; uddni, 2 ; A'dhani, 9 ; kshd'mani ; rdjani ; 
pirshdni, 2. 

2. The locatives in -an (203, from 45 stems). 

A. Masculine: ddhvan, 6. Neuters: anarvdn; aparv&n; dhdnvan,2; dh&n- 
van, 5 ; from AV., dhdnvan, v.13.1 ; p&rvan, xii.3.31. 

B. Masculines: dcman; tmdn, 5; dtmdn, AV. ix.6.21 : xii.3.64: xv.1.2: TS. 
v.1.9 6 : Qat.Br. i.8.1 48 : ii.5.3 4 : iii.1.3 2 '; 4 6 : iv.5.4 5 : v.2.1 5 ; dcman, TS. iv. 6.1 K 

Neuters : djman, 5 ; dydman, 2 ; kdrman, 3 ; kdrman-karman, 2 ; kd'rshman, 
2 ; carman ; j&niman, 4 ; jdnman, 3 ; jdnman-janman, 2 ; ddriman ; dhdrman, 5 ; 
dhd'man, 4 ; pdtman, 6 ; brahman, 2 ; bkd'rman (see above) ; mdnman, 2 ; mdr- 
man, 2; yd'man, 32; vdriman, 5; vdriman; varshman, 5 ; vijd'man; vidharman, 
3; vioman, 12; carman, 17 ; pa.sm.an; sakman; sdd/man,±; sd'man,3; svd'dman; 
hdvtman, 2 ; from AT., kdrman, xii.3.47 ; karman-karman, and yd'man-ydman, 
iv.23.3; dhd'man, iv.25.7: xx.49.3; varshman, iii.4.2 : v.2.7?; vioman, 23 times ; 
from TS., eman, odman, bhdsman, iv.3.1 1 ; brahman, kdrman, 3 2 ; yd'man, 6.1 3 ; of. 
VS. xiii.53 : xvii.1,6. 

0. Masculines: jmdn, 2; pdryman, 5; mfirdhan, 6; mdtaripvan, AV. xi.6.13; 
jmdn, TS. iv.6.1 2 : vii.4.15 1 . 

Neuters: dhan, 15; dsdn, 8; uddn; 0,'dhan, 6; kshd'man ; ptrshan, 2; dsdn, 
AV. xviii.3.11: xix.60.1: TS. v.5.9 s ~ 3 ; kshd'man, TS. iv.6.1 4 ; ytishan, vi.3.11 1 ; 
ahan, Ait. Br. iv.l ; dkshan, Brh.&r.up. ii.3.5: iv.2.2. In composition we find: 
dsdnn-ishu ; nemann-ish; cf. also patir ddn, and rad. r-stems, L.s.n., p. 488. 

Transition-forms. For ptrshb, pHrvdhn'e,, etc., see pp. 526-7. 
Vocative Singulab MAsotrLiNB. 

There is no neuter. For vidharman (v.17.2 — Gr. neut.), cf. 
the masc. nom. vidharmd, AV. xvi.3.2. 

The Vedic voc. of the vant and mant-stems ends regularly in 
-as; that of the van, man, and arc-stems, regularly in -an (cf. p. 
519). Otherwise Benfey, Gram. p. 312, 1. 7. Nevertheless, five 
stems in van (all having feminities in -vari — p. 527) mate voca- 
tives in -vas : rtdvas, 4 ; evaydvas, 2 ; prdtaritvas; mdlari$va$; 
vibhdvas (Gr., stem in vant) ; we may add khidvas, vi.22.4 (BR., 
stem khidvan). 

A. There are 169 forms (from 12 stems) : apvaddvan; rnagh- 
aoan, 145; yuvan ; vamddvan ; satyasatvan ; satrdddvan ; 



V.s.m.] Noun-Inflection in the Veda. 537 

sahasdvan, 13 ; sutapdvan; somapdvan; svaddvan; svadhdvan, 
2; svaydvan; from AV., atharvan, v.11.2; vipvaddvan, iv.32.6. 

B. There are 24 forms (from 11 stems): aryaman, 11 ; purund- 
man; puruhanman; brahman; vidharman; vipvakarman, 2 ; 
vipvasdman; vrshakarman, 2 ; satyakarman; suddman, 2 ; sva- 
dharman; from AV., trindman, vi.74.3 ; sahasradhdman, iv.18.4; 
pdpman, vi.26.1,2; 113.2; suydman, xvLI.1 ; takman, 11 times 
(i.25 : v.22) ; apman, TS. iv.O.'l 1 . 

C. There are 105 forms (from 4 stems): parijman; ptishan, 32; 
rdjan, 46 ; vrshan, 26 ; from AV., urujman, vi.4.3 ; prthujman, 
v. 1.5. 

Nominative, Accusative, and "Vocative Dual Masculine. 

I. There are 121 forms in -d. 

1. The penult is long. A. There are 32 forms in -vdnd (from 
11 stems). B. There are 4 forms in -mdnd (from 4 stems). 
C. There are 19 forms in -dnd (from 4 stems). 

A. The forms are (including vocatives): agrddvdnd; adruh- 
vdnd; rtd'vdnd, 1 ; grd'vdnd; prdtaryd'vdnd, 2 ; maghdvdnd, 5 ; 
yhvdnd, 10; yH'nd (see below); rathayd'vdnd; pubhraydvdnd; 
prushtivd'nd; sajitvdnd. 

B. The forms are: brahmd'nd; satyadharmdnd; sdndmdnd; 
mpdrmdnd. 

C. The forms are : pdrijmdnd; mltr&rdjdnd (v.62.3) ; rd'jdnd, 
16 ; pvd'nd. 

The scansion of the passages in which these 121 forms occur 
shows that the metre demands or favors a short & as penultimate 
in the cadence of three octosyllabic pddas : v. 70.2a: viii.26.19c/ 
38.2a. 

2. The penult is short. There are 66 forms (from 6 stems). 
The metre favors d in the second place of i.10.36: ii.16.5c, 6b: 
x.66.6c. 

The forms are : aryamd'nd; j'em&nd; mdrdpHshd'nd (and AV. 
vi.3.1) ; ptishd'nd; vrshdnd, 60; somdptishdnd, 2; vr'sh&nd, AV. 
vii. 73.1,2; 110.2. 

II. There are 22 forms in -au. 

1. The penult is long. Forms: (A) rtd'vdnau, 2; from AV, 
grd'vdnau, xi.1.9,10; (C) rd'jdnau, RV.,*4; pud'nau, 2 (and AV. 
viii.1.9) ; vr'shdnau, AV. xix.13.1. The AV, at xviii.1.54, 
modernizes rd'jdnd of the Rik (x.14.7) to rd'jdnau. 

2. The penult is short. Forms: (C) vrshdnau, 13; somd- 
pushdnau. 

Weak form for strong. For yil'nd (=zyHvdnd), see p. 523. 

Nominative and Accusative Dual Neuter (see pp. 524^5). 

There is no case of syncopation in the Rik; cf. p. 524 top. 
B. The forms are : edrmani; jdnmani, 2 ; sddmant;— (I.) 
dhd'mani; — (II. b) from AV., nd'm(a)ni, iv.9.10. 



538 C. R. Lanman, [van, man, an-stems. 

C. The forms are: (I.) dhani, 7; from AV., dhani, xiii.2.3 ; 
cdkshani, x.2.6 ; doshdni, ix.7.7 : x.9.19. 
Transition-form : $irshk,, see p. 527. 

Instrumental and Dative Dual Masculine. 

Instrumental : vr shabhydm, 2 ; grd'vabhydm, AV. vi.138.2. 
Dative : somdp'dshdbhydm. 

Gen. and Loo. Dual Masculine and Neuter (see pp. 624-5). 

Genitives masc. : (I.) pdrijmanos; — (III.) indrdpHshn6s. Gr. 
proposes (II. b) maghdrv(a)nrOS, at the end of v.86.36; but the 
catalectic cadence of the text, maghdnos, is better. Genitives 
neut. : sddmanos; — (HI.) dhnos, AV. xvi.8.22. 

Locative masc. : dpmanos. 

Nominative and Vocative Plural Masculine. 

I. The penult is long. A. There are 77 forms in -vdnas 
(from 22 stems). B. There are 34 forms in -mdnas (from 20 
stems). C. There are 18 forms in -dnas (from 4 stems). 

A. The forms are: akshnayd'vdnas ; dtharvdnas, 2 ; anarvd'nas, 2; ayajvdnas, 
2; upahdsvdnas ; r'hvdnas/2; rtd'vdnas, 10; grd'vdnas, 19 ; drihvdrjas ; nikr't- 
vdnas; parishddvdniu ; prasthdvdnas ; prdtwyd'vdnas ; magh&vdnas, 14; y&vd- 
nas, 11; raghwpatvdnas ; vanivdnas; fubhamyd'vdnas ; frushttvd'nas, 2 ; satodnas ; 
sudhdnvdnas ; supukvdnas. 

B. The forms are : addmd'nas, 2 ; adhvasmd'nas ; dbrahmdnas ; arafmd'nas ; 
Shabrahmdnas ; ddmd'nas ; dvijdnmdnas ; dharmd'nas ; dhvasmd'nas ; brah- 
md'nas, 9 ; ' Vhrd'jajjanmdnas ; mahimd'nas, 5 ; glmjcmm&nas ; satyadharmdnas ; 
sukarmdnas ; mjdnimdnas ; sufarmdnas, 2 ; sthd'rapmdnas ; svddm&'nas. For 
jarimd'nas, x.27.21, see p. 534. 

C. The forms are : ti,rdhvdgrdvdnas ; UqmaimtiirdMmas ; m&rdhA'ws,'!,; rd'jdnas, 
14; f rom AV., pdnthdnas (p. 441) ; fvd'nas, iv.36.6 : xi.2.11 ; 10.23. 

The scansion of the passages in which the above 129 forms 
occur shows that the metre favors short & in the cadence of 
vi.46.11e? (and viii.20.los ?). Cf. Kuhn, Beitrdge, iii.l 21. 

I. a. From the stems that show both & and 4 in the strong 
cases occur: ukshdnas, Katy. Cr. xxiii.4. 7,8,11 ; vr'shdnas, Cat. 
Br. xiii.3.3'; takshdnas, Schol. to VS. xvi.27. 

II. The penult is short. There are 64 forms (from 4 stems). 
The metre favors d in the second place of iii.27.155: viii.33.11a: 

x.66.65. 

The forms are: aryamd'nas; uhshd'nas, 12; rbhukshdnas, 12; 
vr'shdnas, 39 ; ukshd'nas, AV. iv.24.4. For pdnthdnas, see p. 441. 
' Weak form for strong. For hr'tvano (AV. xix.35.5) and rnagh- 
6nas (RV. vi.44.12), see p. 523. 

Nominative and Accusative Plural Neuter. 
The vowel-stems have -d, -dni; 4, 4ni; -tf,, -Uni; see pp. 347, 
394, 415. The forms in 4 and -« would appear to be simply 
shortened from 4 and -4; but in fact, for the great majority of 
the latter forms, the pada has 4 or -$,. Both texts agree in having 
long 4 only in the words terMhmt, ptici, and trt' ; in no case do 



N.A.p.n.] Noun- Inflection in the Veda. 539 

both texts have long -4. The pada and Prat, regard the long 4 
and -# of the samhitd as a lengthening of 4 and •■&. Perhaps, 
however, it is better to set aside their authority, in view of the 
analogy of the forms in -d from a-stems (whose long -d is surely- 
organic), and to consider the i and #-forms as original, and the 
2 and &-forms as secondary. 

For the consonant stems, we should at first thought set up the 
inverted series -dni, -d; and regard -d as a shortened form of -d 
(cf. Benfey, Gram. p. 306, 1. 21). But in fact, the intermediate 
long -d is here equally uncertain. The only words that have -d 
in both texts are dhd and pirshd', and for both of these the transi- 
tion-stems in -d are established (dhdndm, pirshd-m) ! Of the 27 
forms that have -d in the samhitd (p. -&), 18 might be explained 
metrically ; but the other 9 can not : the Pr&t., however, accounts 
for the other 9 as lengthenings of a short -d. But here, as before, 
the samhitd-forms are perhaps, after all, the original ones, relics 
of the organic form in long -d; and the forms in -d, the secondary 
ones. It is hard to see why they should all be short in the pada- 
text, and the question is at least an open one. 





A 


B 


C 


Sum 


I. The regular form is -dni : 


9 


139 


18 


166 


II. Both texts have -S, : 


6 


57 




63 


III. The pada has -&; samhitd, -d: 




27 




27 



I. The scansion of the passages in which the 166 forms occur 
shows that the metre demands a short in the penultimate of 4 
octosyllabic pddas : v.5.105; vii.55.6c. - viii.87.85 (trochaic?): 
x.97.2a {dhd'mdni). Cf. Kuhn, Beitrdge, iii.121. 

A. The forms are : dhdnvdni, 3 ; dhdnvdni, 4 ; pdrvdni, 2 ; 
from AV., dhdnvdni, vii.41.1 ; pdrvdni, i. 11.1 : xi.8.12 : xii.5.71 : 
snd'vdni, xii.5.69. 

B. The forms are : kdrmdni, 8 ; edrmdni; jdnimdni, 9 ; jdn- 
mdni, 3 ; dhdrmdni, 8 ; dhd'mdni, 25 ; nd'mdni, 14 ; brdhmdni, 
53 ; mdnmdni, 5 ; mdrrndni; rdmdni, 3 ; vdrtmdni; pdrmdni; 
sddmdni, 3 ; sd'mdni, 4 ; from AV., dd'mdni, Umdni, vdrmdni, 
etc. 

C. The forms are : akshd'ni; dhdni, 14 ; pirshd'ni, 2 ; sakth- 
d'ni; from AV., dhdni, vii.69.1 ; 82.4,5 : xix.7.1 ; pirshd'ni, i.7.7: 
x.4.19. In RV. i.l40.13c, the metre demands the longer gram- 
matical form dhdni, text dhd. 

II. Both texts have -d. In i. 173.3a, we have sddmd (certainly 
plural), with short final d as eighth syllable in a pdda of 11, 
where it ought surely to be long. 

A. The forms are : dhdnva; dhdnua; pdrva, 4. 

B. The forms are : karma, 5 ; jdnima, 2 ; jdnma, 3 ; dhdrma; 
dhd' ma, 7 ; nd'ma, 13 ; brdhma, 8 ; bh-a'ma, 3 ; mdnma, 3 ; rdma 
(or sing.) ; pdrma, 7 ; sddma, 4. 

III. The samhitd has -d; the pada, -d. Here belong 27 forms 
from maw-stems only. 

B. The forms are : jdnimd, 9 ; dhdrmd, 2 ; nd'md; brdhmd 
6 ; bhU'md, 6 and in vi.62.8 (? Wb. 1765) ; r6md; sddmd. 



540 C. R. Lanman, [van, man, an-stems. 

In 8 instances (j&nimd, 6 ; dharmd; nd'md), the -d is eighth syllable in apdda 
of 11 or 12 (Prat, viii.21); in 10 instances (brahmd, 6; bh&'md, 3 ; romd), the 
second in 5, 8, or 11 (Prat, vii.25: viii.14): in these, the -d might be metrical. 
On the other hand, in 6 instances (dharmd ; ih&'md, 4 ; sadmd), the -d is at the 
end of apdda (Prat, viii.30: vii.14); and in 3 instances (janimd, 3), the seventh 
in 11 (Prat, viii.5). Here the -d is not simply metrical. 

The AT. has: janimd, p. -&, ii.28.2: iv.1.3 : v.11.5; nd'md, p. -&, v.1.3 ; 
vdrimd, p. ?, iv.25.2. In vi.31.3 : ix.9.3 : xix.58.4 (=RV. x.189.3 : i.164.3 : 
x.101.8), the AV. has dhd'md, nd'md, vdrmd, where the RV. has -#. 

IV. Both texts have -d. The Rik has but two such forms : 
firshd', 4 ; dhd, 21. Both may be referred to vowel stems. The 
AV. has : pdrvd, p. ?, xii.5.42 ; dhd, v.1.1 : xx.135.7. 

Accusative Plural Masculine (see pp. 524-5). 

A. The forms are: ddhvanas, 11; dyajvanas, 2; from A V., 
dr&hvanas (iv.29.1), etc.; — (I.) maghdnas, RV., 10; — (II. b) 
drdv(a)nas, 3 ; (II. c) drdv(a)nas; evayd'v(a)nas; — (III.) grd'v- 
nas, 2; prdtaryd'vnas; evayd'vnasf, text -yd's, p. 451 end; — 
yH'nas, 2. For yuvdtas, see p. 516. 

B. The forms are : dpmanas; brahmdnas; takmdnas, AV. 
v.30.16; — (HI-) from AV., durnd'mnas, xvi.6.7 : xix.36.4,5. 

C. The forms are : mdtaripvanas, AV. xix.27.4 ; — (I.) uksh- 
dnas, RV. x.86.13 : so even in Bhag. P. x.83.13, BR. ; vr'shanas; 
— (II. a) vr'sh(a)nas, viii.7.33 ; — (HI.) ukshnds, 2 : so Katy. Qr. 
xxiii. 4.9,10 ; janard'jnas ; mUrdhnds ; yamdrdjnas ; rd'jnas ; 
vr'shnas, 2 ; from AV, majynds, ii.12.7 : ix.5.23 ; 8.18 ;— finally 
punas, RV, 2, and AV. iv.5.2. 

Strong form for weak. Benfey, Gram. p. 311, note 3, cites 
mahdtmdnas as an Epic A.p.m. 

Instrumental Plural Masculine and Neuter. 

A. Masculines (48 forms, from 15 stems) : ddhvabhis, 2 ; 
dprayutvabhis ; r'kvabhis, 15; evayd'vabhis ; grd'vabhis, 4; 
prdtaryd'vabhis, 2 ; ydjvabhis, 2 ; yi'ivabhis; rdnvabhis; pikva- 
bhis, 2 ; sdtvabhis, 1 1 ; samtvabhis; sayd'vabhis, 3 ; suprayd'- 
vabhis; svayugvabhis. 

Neuters : pdtvabhis; pdrvabhis; vivdsvabhis; somapdrvabhis. 

B. Masculines: dkhidraydmabhis ; adhvasmdbhis, 3; dpma- 
bhis ; dprnahanmabhis ; dpuMmabhis ; dmabhis; brahmdbhis; 
mayd'raromabhis; vtdupdtmabhis ; sukdrmabhis, 2 ; sumdnma- 
bhis, 2 ; mpdrmabhis. 

Neuters (67 forms, from 17 stems): djmabhis; 'emabhis; Tcdr- 
mabhis, 2 ; trd'mabhis; dhdrmabhis, 7 ; dhd'mabhis, 13 ; nd'ma- 
bhis, 7 ; pdtmabhis; brdhmabhis; bhdrimabhis, 2 ; mdnmabhis, 
13; yd'mabhis, 4; vdrimabhis, 4; vig dmabhis ; pdkmabhis ; 
sd'mabhis, 4 ; hdvimabhis, 4. 

C. Masculines: ukshdbhis, 4 ; rd'jabhis, 5 ; vr'shabhis, 11. 
There is no stem ni-kdm-an. To me it admits of no doubt that 

the form nikdmabhis at the end oHx.Q2.9d is used (instead of the 
proper nikdmais) simply on account of a formal and metrical 



I.p.m.n.] Noun- Inflection in the Veda. 541 

parallelism with evayd'vabhis at the end of c. Cf. rdnvabhis, at 
end of v.44.10c, and article ranvd. 

Neuters : akshdbhis, 8 ; asthdbhis ; dhabhis, 9 ; dsdbhis, 6 ; 
uddbhis, 5 ; 4' dhabhis, 3 ; mahdbhis,' vr'shabhis; pdkabhis, TS. 
v.7.23 1 . 

Similarly, naktdbhis is used instead of ndhtam at the end of 
vii.104.18c to eke out the pdda to 12 syllables, and so make it 
metrically parallel with a, b, and d. 

Transition-form. For dpnais, see p. 527. 

Dative Plural Masculine and Neuter. 

A. Masculines : grd'vabhyas; yi'ivabhyas; dtharvabhyas, Gr., 
-bhias, vi.47.24. 

B. Masculine : brahmdbhyas, 3. Neuter : dhd'mabhya.i, 2. 

C. Masculines: rd'jabhyas, 4 ; pvdbhyas, AV. xi.2.30 ; tdksha- 
bhyas, VS. xvi.27. Neuter : dhabhyas. 

Ablative Plural Masculine and Neuter. 

A. Neuter: snd'vabhyas, AV. ii.33.6. 

B. Neuter : Idmabhyas. 

C. Masculine : majjdbhyas, AV. ii.33.6. Neuter : dhabhyas, 
RV. 

Genitive Plural Masculine and Neuter (see pp. 524-5). 

A. Masculines : anarvdndm, 2 ; dyajvandm; sdtvandm, 2 ; — 
(I.) maghdndrn, 4 ; maghdndm, Gr. -aam, 12 times at the end of 
catalectic pddas of 7 syllables, and in vii.32.7 (pdda of 11); 

maghSnaarn, with resolution, vi.68.2 : viii.1.16 ; 2.34 ? ; 19.34 • 

(II. b) rdrdv(a)ndm ; vdjadd'v(a)ndm ; somapd'v(a)ndm ; 'for 
sahasradd'v(a)ndm, see p. 528 top ;— (III.) grd'vndm, 2. 

Neuter: dhdnvandm, AV. xi.9.1. 

B. Masculines: brahmdndm;— (III.) durnd'mndm, AV.n.25.2. 
Neuters: dhdrmandm, 2 ; brdhmandm,2;'mdnmandm;— (III.) 

dhindmndm; from AV., sd'mndm, xv.6.3. 

0. Masculines: (III.) >vr'shndm, viii.72.1: viii.46.29 (Gr. -a<2»;) ; 
from AV., rd'jndm, thrice ;— fdndm, AV. iii.9.4. 

Neuter: dhndm, 28 times. 

It has teen suggested that mahi'nam, x.134.1, might be G.p.m. of maMman 
the m being dropped as in the instrumental. If that were so, we should have 
maMn&'m (with 5, and oxytone). But cf. pp. 398-9. 

Transition-form. For dhdndm, see p. 527. 

Locative Plural Masculine and Neuter. 

A. Masculines : ddhvasu; kr'tvasu; ydjvasu, 2. Gr. reads 
maghdshu, x.93.14, text maghdvatsu. 

Neuters : dhdnvasu, 7 ; dhdnvasu; pdrvasu. 

B. Masculine: dpmasu, AV. iii.21.1 : xii.1.19. 

Neuters : Jcdrmasu, 2 ; jdnmasu; dhd'masu, 3 ; vdrmasu: sdd- 
masu, 2. ' 

YOi.. x. 72 



542 C. R. Lawman, [m-stems. 

C. Masculines: rd'jasu ; from AV., majj&su, i.11.4 ; yama- 
rdjasu, xviii.2.25. 

Neuters : dhasu; pirshdsu, 2. 

STEMS IN INI 

The stems in in are regularly oxytone. They neither syncopate 
the thematic vowel, nor lengthen it in the G.p., nor throw the 
accent forward to the ending. Their declensional forms present 
almost no irregularities whatever. Anomalously accented are 
irin, sarin and pd'kin (i.51.8 — a mere mistake for $dM ' ?). The 
compounds with negative a- are barytone ; and so kdrildatin, 
mahd'vat-drin, and indramedin. 

Besides the stems formed by the suffix in, are here included 
several in vin (cf. mdydvin with mdyin) and one in min (rgmin). 

These stems are declined only in the masculine and neuter. The 
fem. is formed with i: as, arkin-i, p. 367. 

Transitions to the a-declension. The A.s.m. parame-shthin-am 
is the point of departure for the N.s.n. parame-shthina-m 
(instead of -shthi). The form is especially interesting, because, as 
is perfectly evident, it owes its origin solely to the metre (so 
BR.) : iddm ydt parameshthinam \ mdno vd brdhmasampitam, 
AV. xix.9.4. Corresponding to vanin-am, as if it were vanina-m, 
is formed (instead of vanin-as) the false plural vanindni, also due 
to the metre : x.66.95, d'pa dshadhir vanindni yajniyd. In viii. 
33.6, 1 would venture to suggest the reading pdkina-s, as transi- 
tion-nom.s.m. to pdkin-am. The text has pdMnds {ga-As, as in 
vi.45.22) ; but cf. for the accent am-ind. Possibly mahina (indra, 
vi.26.8) and mahinasya (indrasya, vi.33.5 ; 68.8) are transition- 
forms ; cf. mahin-e (indrdya, vii.31.11). 

Transition to the *-declension. Perhaps the an. X. mandi-m 
(i.9.2) is an A.s. falsely formed after the analogy of the Ls. 
mandin-d, as if it were mandi-nd (cf. p. SIS). BR. explain it 
otherwise ; see A.s.m. 

On the other hand, ni-yay'i-nam may belong in the same cate- 
gory with Armi-nam, khddi-nam, p. 378. 

THE NEUTER FORMS. 

The neuter forms are very rare. They are : dndmi, 2 ; ubhayd- 
hasti; patatri, 4; from AV., pakalyesM, i.25.2 ; tsdri, x.1.19; 
kirin-d, RV. v. 4.10; 40.8; sugandhin-d, viii.19.24 (but cf. p. 
373); prahoslunas, G.s.n., viii.81.4; parameshthinam and van- 
indni, above. 

THE MASCULINE FORMS. 
Nominative Singulab Masculine. 

Here belong 213 forms (from 55 stems). In vii.60.10, Gr. takes 
tveshi' as N.s.m. with sdmrtis; but since this is fem., it is better 
to take tveshi' as N.s.f. from tveshia (cf. p. 369 med.). BR. read 
tveshyS eshdm (a misprint — see vii.1751). In x.51.6, rathi'va 
probably stands for rathi'r-iva; see p. 375 end. 



N.am.] Noun- Inflection in the Veda. 543 

Enumeration : ddvaydvi ; ayudhvi ; iri ; kar&datl ; (d'ki ? ; sari ; — anki' 
abhydvarW ; arthi' ; oqvi', 4; ashtrdvi' ; vdrt', 2; rgmi' ; rjipi' ; rjishi', 10 
kapardi' ; kdmi' ; kevalddi' ; kefi', 4; tsdri' ; dyumni', 10; nishshapi' ; paJcthi' 
paripanthi' ; prshtidmayi' ; prapathi' ; prabhangV ; pravepani' ; bdhufardhi' 
brahmacdri' ; bhadravddi', 2 ; manishi', 5 ; mandi', 6 ; mardyi' ; mahdhasti' 
mdyi', 3; medV ; rafhi', 3; rarribM' ; rukmi' ; vajri', 21; varmi', 2; vagi', 9 
v&ji', 70; vwaptf,', 3; fatasvi' ; pipri', 4; fushmi', 10; fhJtf ; fvaghni', 6 
sahasri', 2 ; sdci' ; somi', 2 ; svabdi' ; hastV. 

Accusative Singular Masculine. 

There are 192 forms (from 53 stems). For ■drtni-nam (ix.98.6), 
and Middi-nam. (vi.16.40), see «'-stems, p. 378. The latter word 
yields a tolerable sense even if referred to an m-stem (Jchddin-am) ; 
but in that case, I would suggest the reading hdstena for haste 
net: 'whom they carry (with) on the hand as a new-born child, 
decked with spangles.' Like Urmv-nam is perhaps ni-yayirnam, 
x.60.2 ; cf. yay'irna. Kuhn, Beitrage, iii.46'7, has observed that 
the cadence of i.85.3c requires a long ultima in abhimdtin-dm, 
text -dm; cf. vidh&vdm, p. 330. 

BR. regard mandim (i.9.2) as an inaccurate rendering of the 
pronunciation of mandinam (in-), after elision and crasis; but it 
may be a transition-form (p. 542). The art. A. rjishdm (i.32.6) 
may be explained without setting up a separate stem rjishd ; 
namely, it may have been used instead of rjishinam, in order to 
satisfy better the metrical parallelism of the different pddas and 
the formal parallelism of pdda b (cf. viii.65.5). 

Enumeration: advaydvinam, 2; dvdjinam ; — al/rmaim, 11; ohhimdtinam , 
arkinam; avakrakshinam ; afvinam, 12; dddrinam ; ishminam; vMka$ahsinam ; 
ukthinam, 2 ; udrinam, 5 ; vhhay&vinam ; rjishinam, 4 ; kapardinam, 2 ; karara- 
bhinam, 2 ; kalmalikinam ; kdminam ; kdrinam, 2 ; kvM/yinam ; k&eidarthinam ; 
catiTiam ; tarasvinarn, ; dyumninam ; dhanitiam, 2 ; tiamasvvna/m ; niyayinam ; 
pakshinam ; paripanthinam ; putrinam ; purishinam, 2 ; manishinam, 2 ; manth- 
inam ; mandinam, 3 ; manyusjidvinam ; mdydvinam ; mdyinam, 10 ; rakshas- 
vinam ; rathinam ; vajrinam, 6 ; vaninam, 2 ; vaninam ; vaydkinam ; varcinam ; 
vdjinam, 49 ; virappinam, 2 ; fatagvinam, 4 ; faiinam, 6 ; fdkinam, 2 ; fushmi- 
nam, 3; frnginam; saJiasraposhinam ; sdhasrinam, 27. 

Instrumental Singular Masculine. 

Here belong : mahd'tiaMrind; — arcind; apoind; Mmidmd, 2 ; 
matidind, 2 ; rakshasvind ; vattbrind ; vdjind, 3. For kirind, 
i. 100.9, Gr. proposes -ne. 

Dative Singular Masculine. 

Here belong : abhydvartine; amitrine; kapardine, 2 ; kdrine; 
kimidine ; prasakshine ; bhdmine ; manisMne ; mandine, 2 ; 
mahine; rakshasvine, 2; rathine; vajrlne, 12; viraptfine; pdkine, 
2 ; ciprine, 2 ; cushmine, 2. 

Ablative and Genitive Singular Masculine. 

Ablatives: abhiprabhanginas; patatrinas, 2 ; sominas (x.57.1). 

Genitives (71 forms, from 27 stems) : ddvaydvinas; — arthinas, 

2 ; rjishhnas; dvaydvinas; dhaninas; namasv'mas, 2 ; parninas; 



544 G. R. Lanman, [m-stems. 

pragardhinas; bh/driposhinas ; mandinas; mdyinas, 7 ; ratninas; 
rasinas, 3 ; retinas; vajrinas, 3 ; vaninas; variinas, 2 ; varcinas, 
3 ; vdjinas, 17 ; virappinas, 2 ; viduharshinas ; vraudinas ; 
patinas; pushminas, 6; pragmas; sahasrinas, 2; sominas, 7. 

Locative Singular Masculine. 

Here belong : mdyini; vdj'mi; som'tni; from A V., dvipini, 
hastini, vi.38.2; parameshthini, x.3.24; brahmacdrini, xi.5.22. 

Vocative Singular Masculine. 

Here belong 76 forms (from 1 5 stems) : dndbhayin; amatrin; 
ubhaydvin; rjlpin; rjishin, 9 ; tuvikHrmin; prapathin; yakshin; 
vajrin, 41 ; vdjin, 5 ; virappin, 2 ; pavasin; piprin, 6 ; pushmin, 
4 ; sahasin. 

Nominative, Accusative, and Vocative Dual Masculine. 

I. There are 369 d-forms (from 15 stems) : according to Mailer's 
Index, apvind, 68 ; apvind, 246 ; -d, p. -au, 4 ; dpvind, 27 ; — 
kumdrind; kepind, 7 ; pajrahoshind; parnind; putrind; puri- 
shind; prasakshind; raanthind; mdydvind; mdyind; vajrind; 
vdjind, 4; vdjind; vighanind; pushmind. The A V. has: apvind, 
14; apvind, 17; dpvind, 4. 

II. There are 32 (m-forms : apv'mau, 11 ; apvinau, 17 ; dpvinau, 
2 ; vdjinau, 2. 

In the very late verse x.184.2, the Rik reads, gdrbharh te apvinau devd'v a' dhat- 
tdm pushkarasrajd ; and the incorrect AV. variant (v.25. 3), . . apvinobhd' dhatt&m 
. . The Rik-verse, if conformed to the canons of the old parts of the Eiksanhita, 
would read : apvind devd' dhattdm, p. devd' d' ; or apvind ubhd', p. -nam ubhd' d'. 
The Atharvan requires emendation. We may read : apvindv ubMfv d' ; or, -nau 
devd'v d'. 

Inst., Dat., Gen., and Loc. Dual Masculine. 

Instrumentals : apvibhydm, i.44.2,14; 182.6: v.51.8 ; indra- 
medibhy dm, AV. xi.9 .4. Dative: apvibhydm, i.164.27 : v.73.10. 
Genitive: apvinos, 17. Locative: dpvinos, i.120.10. 

Nominative and Vocative Plural Masculine. 

Here belong 153 forms (from 64 stems). Some may be singular. 

Enumeration : ankupinas ; atrinas, 2 ; arkinas, 2 ; wcinas, 2 ; arthinas, 3 
ishminas, 3 ; ukthinas, 6 ; upamantrinas ; rgminas ; rjishinas, 2 ; kapardinas 
kabaridhinas ; kdminas, 2 ; kdrinas, 2 ; kirinas ; krpaninas ; kepinas, 2 ; khddinas 
gdthinas ; gdyairinas ; gharminas ; tarasvinas, or G. ; dapagvinas ; drapsinas 
dvaydvinas, 2 ; dhuminas ; dhrshadoinas ; namasvinas, 5 ; nitodinas ; nirdminas , 
nishanginas, 2 ; pakshinas, 2 : patatrinas, 2 ; paripanthinas ; parninas, 2 ; pdpinas, 
2; purisMnas ; prakridimas ; prasakshinas ; balinas ; manishiiias, 16; voc, 5 ; 
mandinas, 5; mdhinas; mdydvinas ; mdyinas, 4; medinas; rathinas; vaninas,2- 
vaninas, 4 ; varminas ; vdjinas, 16 ; voc, 2 ; virappinas, 2 ; voc, 1 ; virokinas, 2 , 
vishdninas ; vratacarinas ; patagvinas ; patinas, 3 ; pdkinas ; pushminas ; sahas 
rinas, 4 ; sominas, 5, see Ab. ; voc, 1 ; svdninas ; Jiastinas, 3 ; Mraninas. 



A.p.m.] Noun- Inflection in the Veda. 545 

Accusative Plueal Masculine. 
Here belong 44 forms (from 23 stems). 

.Enumeration : atrinas, 3 ; ardjinas ; a/rihinas ; afvinas ; ukthafansinas ; 
kdminas ; kdrinas ; grathinas ; dyumninas ; pakshinas ; bhdminas ; mandinas ; 
mdyinas, 3 ; mitrinas, 2 ; rakshasvlnas, 3 ; rathinas, 3 ; vaninas ; vanlnas, 7 ; 
vdjinas, 5 ; vrandinas ; sahasrinas, 3 ; sominas ; Mraninas. 

Instrumental Plueal Masculine. 

Here belong : apvibhis; rgm'ibhis; kepibhis, 2 ; nishangibhis; 
patatr'ibhis, 3 ; manishibhis, 1 ; mandibhis; rukmibhis; vdjibhis, 
4 ; pushmibhis; hastibhis. 

Dative Plural Masculine. 
Sole example : pikhibhyas, AV. xix.22.15 (? MSS. pishir). 

Genitive Plural Masculine. 

Here belong : kdrl'ndm; mdyi'ndm, 3 ; vdjindm; prngindm; 
stukdvmdm; vaplndm, AV. xv.11.9. Cf. p. 373, IV. 

Locative Plural Masculine. 
Here belong: Jchddishu; dvarishu; hastishu, AV. xii.1.25. 

STEMS IN AS. 

The words whose stems end in as are primarily and principally 
neuter nouns, accented on the radical syllable : as, mdn-as. 

They often occur, however, at the end of compound adjectives, 
and are then inflected in all three genders : thus, N.s., su-mdnds, 
m. ; su-mdnds, f. ; su-mdnds, n. There are also a few oxytone 
masculines in as : as, rakshds. 

Of a few stems, the formation is not clear : upas (? p. 496) ; 
ripd'das. For Jchi&vas, see p. 536. The stem r'bhvan is sup- 
plemented by r'bhvas and r'bhva; and p'ikvas, by pikvan and 
pikvd — see p. 526. Gr. sets up a stem ush=.ushds; but see Ab.s. 
Cf., however, d's, tid, up, ytl's (p. 523), and dksh (? see G. dual). 

For the accent of bhish-d', see I.s.m. 

Stems in as distinguish strong and weak cases only in the 
N.A.p. neuter, where the thematic a is lengthened and nasalized. 

Benfey (Gram. p. 316, xvii.) admits the existence of strength- 
ened forms in the masculine and feminine. The contract forms, 
however, can be accounted for without the assumption of a long 
vowel (-dm=z-dsam, not -dsam) ; and as for the other apparent 
instances of strengthening, they are all more or less doubtful. 

The least doubtful case is that of the an. X. topd'sd (rathayd 1 - 
vdnd), viii.38.2, p. topd'sd. This is the only instance in which 
both texts have d. Brugman, Zeitsch. xxiv.24, has suggested 
topd' s&rathayd'vdnd. 

The most numerous are the much-discussed forms of ushds. 
The strong forms with short 4 number 78 : A.s. (ushd'sam), 33 ; 



546 G. B. Lanman, [as-stems. 

du., 4 ; JsT.V.p., 41. Over against these are to be set 28 strong 
forms with long d in the samhitd, and in the samhitd only : A.s. 
(ushd'sam), 11 ; du., 3 ; N.V.p., 14. It is most important to 
observe that in not a single instance does the padapdtha have df 
and further that the d is imperatively demanded by the metre in 
20 instances, and decidedly favored by it in all the rest. 

In 18 instances the d is penultimate of a pdda of 11 syllables: i.32.4; 123.12; 
124.9,13: ii.28.9: iii.31.4: iv.3.11; 5.13; 12.2: v.1.1: vi.30.5 ; 65.4; 72.2: 
vii.41.7 ; 42.5 ; 72.4 ; 76.4; 88.4 The d occurs as eighth syllable, in a pdda of 
11, at vii.99.4; of 12, at x.35.2. The d occurs as second syllable in a pdda of 
8, at i.188.6 : iv.30.9 : viii.27.26. It occurs as fourth syllable, in a pdda of 8, at 
v.5.6; of 11, at viii.85.1 : x.88.18; of 12, at i.92.2 ; 134.4a. The compound 
ushd'sd-nakta occurs ten times, always at the beginning of a pdda of 11 or 12, so 
that the d is in the second place. Naktoshd'sd (5 times) also occurs only at the 
beginning of a pdda of 8 or 1 1 ; so that we find o in the third place, despite the 
metre. 

Brugman. l.c. 21, actually divides the masculines and feminines into two 
classes, of which one distinguished the strong cases from the weak, and the other 
did not. His first category contains only one stem, ihshas (p. 28). But the 
evidence of the Vedic forms upon which his argument is based seems to me too 
equivocal to have much weight. It is far from certain that the whole phenomenon 
is not a purely metrical one; cf. vii.41.6a with 7a: 78.46 with 76. id: i.123. 11a" 
with \2d. I examined the 78 passages in which the strong cases with a occur 
and found not one where the metre demanded d, and only three where it even 
favored a: viz., iii.55.1a, a late hymn, a second in pdda of 11 ; vii.75.36, 8, fourth 
in 11 ; x.127.36, a second in 8. 

The long d of a-yd's is inherent in the stem ; ci also ayd'sya, dydsd. Nd's-d 
is a strong form of nets ; but this is a radical stem (p. 493). 

It is by no means certain that the an. 1. in i.l68.9c, te sapsard'so 'janayanta 
dbhvam, is to be referred to an as-stem ; but if so, the metre is almost enough 
to justify the long d. BR. and Gr. refer it to sapsara-. 

For vd'ta-svanas-as, vii.56.3c, we ought to read vd'tasvandsas (p. 345) since an 
aksharapankti demands a long in the fourth place. But this form we refer of 
course to the a-stem (cf. vd'ta-svana-m, viii.91.5). 

We also find one weak case with d: viz., in x.39.15, doshd'm 
ushd'so hdvio havishmatd, p. ushti'sah. Gr., Wb. IT 51, takes it 
as G.s. ; it may be A.p. As for ushd's-dm (cited by Benfey, SV. 
Gl. 32,150, and Gram. p. 316), it is indeed favored by the metre, 
RV. viii.43.5c, ushd'sdm iva ketdvah, but the text has <t'. 

The Zend has apparently strengthened forms. The A.s. ushd- 
onhem occurs 5 times. In Vend, xiii.l, all the MSS. have vtppem 
(! masc.) paiti ushdonhem, so that one is tempted to ask whether 
the word has not here gone over to the a-declension — ushdsa-m; 
see below, III. The N.p.rn. ha-zaoshdonh-6, i. e. sa-jdshds-as, 
occurs Yac. li.20. But we also find a weak case with strength- 
ening, viz., the A.p. masc. (yazamaidi vaMscd) awezh-ddndonh-6, 
i. e. *ab-ddnds-as, Yac. xlii.2. It is certainly no easier to take 
vairis as fern., and make -ddndohhd an A.p.f. of -ddna (cf. p. 363 
end). 

THE RELATION OF EQUIVALENT STEMS IN iS AND A (A). 
In looking through the Vedic vocabulary, one cannot fail to 
notice a large number of stems in as, by the side of which are 
found stems almost or entirely like them, barring the final s. 



Transitions.] Noun- Inflection in the Veda. 547 

Some of these pairs consist of words of like meaning, but of 
independent origin or formation ; as, mr'gaya and mrgayds (see 
Gr.). Others are similar in formation, but the meanings are 
different ; as, chdnda and chdndas. Tbere are yet other couples 
of true synonyms, of which both are very common (cf. drdvina 
with drdvinas) or equally rare (cf. durdsha with duroshas; Ivdra, 
' circuit, Umkreis,' with vdras, evpos) : and of such it is hard to 
say that the one is less original or normal than the other. 

After leaving all these out of view, there still remain many 
stems in as identical in meaning with stems in a (cf. Benfey, 
Gram. p. 316, note 4). 

If we examine the usage of the language respecting these 
doublets, we find that in many instances the one stem is decidedly 
regular and common, and the other only of sporadic occurrence 
or often even a anaS Xeyo/ASvov, i. e. comparatively irregular. 
In such cases, the right to set up two independent stems, one in 
as and one in a, may fairly be questioned. Upon this subject 
Professor Benfey, Abh. xvii.22, remarks : " These are irregularities 
which are not to be disposed of by inventing a new stem for 
every irregular form. On the contrary, we must endeavor to 
discern how the original declension of certain themes was so 
transformed by heteroklisis and metaplasm, that not only single 
cases assumed two or more forms, but also their whole declen- 
sional system became doubled and manifold." 

In the following pages, therefore, I have attempted to discover 
the relations between the a and the as-stems, in cases where the 
one seemed primitive and normal, and the other secondary or 
derivative, and to trace the processes by which the latter came 
into use beside the former. I find that where any significant 
relation exists it is generally one of transition ; i. e., the secondary 
form is based upon the primary, and is made in accordance with 
the ordinary norms and analogies of the language, but these are 
false. 

Thus the N.s. of the very old word for ' moon, month' is md's; 
the A.s., md's-am; Lb., mds-d'; etc. This remained in vogue as 
a name for ' month ;' but went out of use at an early date as a 
name for ' moon.' In the latter signification it was replaced by a 
comparatively modern word, candrd-mds (d). The recession of 
the accent helped on the oblivion of its etymology ; and at last it 
was no longer felt as a compound of md's. Its N.s. was regular 
— candrdmds. Its A.s. should be candrdmdsam. But in the 
vast majority of instances, the A.s. which corresponds to a N.s. 
in -as ends in -ds-am. Here, then, in accordance with the false 
analogies of the language, an A.s. candrdmds-am was formed to 
correspond to the common N.s. candrdmds, as if this were from 
a stem candrdmds. The other cases were made in like manner, 
and so the whole declension was transformed. Cf. below, IV. 1. 
The N.s. is the point of contact of the two declensional series. 

It is not unworthy of notice that these processes of change 
were often occasioned or aided by the exigency of the metre, or 



548 0. R. Lanman, [as-stems. 

of the formal parallelism of the verse, or by some other circum- 
stance. Of these circumstances I now give a few typical exam- 
ples, and shall notice others here and there in the progress of the 
discussion. 

It is fair to assume that the Vedic poet would not depart from the ordinary 
usages of the language without some reason. And if, upon examination, we find 
these grounds such as give us a less exalted opinion of the general character of 
much of the Vedic poetry, we ought not on that account to doubt their validity. 

The choice between two equivalent case-ending s is often determined by the 
metre ; cf. p. 535 med. After what Aufrecht has said (Rigveda?, preface, pp. 
xii-xxxvii), it is clear that even the words of stock-phrases were altered on the 
same account. A comparison of ii.34.6 with x.42.7 illustrates this. In the 
former, pddas a, b, and c are of 1 2 syllables each : in the latter, of 1 1 : and the 
fourth pddas read accordingly 

kdrtd dhiyarh jaritre vd'japecasam, and 
krdhi' dhiyam jaritre vd'jaratndm. 

Since words of different metrical value, but of equivalent meaning, could be 
gained by using an as-stem instead of an a-stem, a word whose theme is almost 
invariably vocalic (e. g. posha) is sometimes used as if it were consonantal, and 
vice versa. Thus it admits of no doubt to my mind that the form (rdyirh) 
vicvd'yuposhas-am was used in i.19.96 and vi.59.9d rather than -posha-m, for no 
deeper reason than to eke out the pdda to 8 syllables. Since the raison d'Stre of 
these isolated forms is so clear, we are not justified in setting up a stem -poshas, 
except for lexical convenience. 

The stock -pddas ix.68.10c and x.45.12c read: adveshe dyd'vdprthivi' huvema ; 
and a, t, and d are each of 1] syllables. Had they been of 12, c would almost 
infallibly have read huvema dyd'vdprthivi' adveshdsd. 

Forms from a stem dngira were never current and in vogue. And yet we find 
the isolated L.s. dngire at iv.51.4. I believe it is simply due to the formal 
parallelism of the verse : yend ndvagve dngire ddfagve saptd'sie, etc. 

In ii.10.3, bhuvad agnih purupecdsu gdrbhah, the form purupecassu would have 
satisfied the metre (ss would make the eighth syllable long by position — see 
L.p.n.), but would have been of ambiguous gender (m.f.n.). I believe the d-form 
purupecdsu is used in order to distinguish the word as a feminine (sc. mdtr'shu — 
Say., oshadhishu). 

For the more systematic treatment of the subject in hand, I 
shall dispose it under five or six several heads. 

A. 

I. Transitions from the declension in as to that in 4 (fem.), 
and a (masc), by contract forms. 

II. Other transitions from the as-declension yielding forms 
from shorter vocalic themes. 

III. Transitions from the as-declension resulting in forma from 
longer vocalic themes. 

B. 

IV. Transitions of radical stems to the os-declension. 

V. Transitions of suffixal a-stems to the as-declension. 

VI. Miscellaneous cases. 

The following table shows some of the principal points of con- 
tact of the two declensional systems. 



Tk 


msitions.] 


Noun- Inflection in the 


Veda. 






The 


makes 




This last form 


of a 




stem 


its 




is regarded as 


stem 


i. 


dfas 


A.s.f. 


Ofd'm. 


A.s.f. 


dfd'-m 


dfd'. 




ndvedas 


N.p.m. 


ndvedds. 


N.p.m. 


ndvedds 


naved&. 


ii. 


hedas 


N.s.n. 


hedas. 


N.s.m. 


heda-s 


heda. 


IIT. 


■tejas 


A.s.m. 


-tejas-am. 


A.s.n. 


■tejasa-m 


-tejasa. 


IV. 


-dhd" 


N.s.m. 


-dhd'-s. 


N.s.m. 


■dhd's 


■dhti's. 


V. 


(epa 


N.s.m. 


fepa-s. 


N.s.n. 


fepas 


i&pas. 



549 



I. Transitions from the ^-Declension. Contract Forms. 

The subject of contract forms is so inseparably connected with 
that of transition to the vowel declension that both must be dis- 
cussed together. 

The existence of contract forms from «s-stems is mentioned by 
Benfey, Gram. p. 316, 1. 14. Gr., Ueb. ii.518, remarks that they 
are indeed infrequent, but not unheard-of. Benfey derives the 
contract forms in -dm and -ds from strengthened forms in -dsam 
and -dsas. This seems unnecessary. Forms in -d(s)atn, -d(s)as 
would yield the same result. Su-mdnds-am becomes su-mdndm; 
and su-mdnds-as, su-mdnds. Just so in Greek sv-jxsvicf-a^fA.) 
becomes Ev-p.£vi) ,• and sv-^sveff-ei, sv-fxeveli. 

The masc. stem vedhds is well authenticated ; cf. vedhds-am, 
v.43.12; vedhd's (sdmds), ix.101.15. But in ix.102.4, we have 
vedhd'm (jajndnd'm, sc. sdmam), and in ix.26.3, vedhd'm (td'm). 
There can be no reasonable doubt tbat this is a true contract 
form. The assumption of a masc. stem in d is out of the question. 

These contract forms are not unknown to the Avesta. So on 
p. 330, 1. 8, ed. Westergaard, we find ushomi priram. In Yacna 
lxii.5, occurs frazantim hvdpdm, equivalent to prajd'm sv-dpdm, 
i. e. su-dpasam; the norn. is hvdpdo, i. e. su-dpds. In Yc. li.18, 
qarendo is probably an ablative singular of the stem qarenanh, 
for qarenanhd, like the Vedic G.s. nrmdnds for nrmdnas-as. Of. 
also Justi, Ilandbuch, § 406. Mazddo as N.s. (cf. medhd's) is 
dissyllabic ; but as G.s. (cf. medhds-as), trisyllabic. The A.s. 
mazdam still retains traces of its trisyllabic origin (cf. medhds- 
am) in the Gathas, e. g. in the octosyllabic pdda, Yc. xxviii.4.2a, 
mazdamcd ahurern; but not in the Yashts, e. g. yd mazdam tard 
manyttt, x.139. See Mayr, Sitzungsberichte d. kais. Ah. (Vienna, 
1871) lxviii.774, and Geldner, Metrik, p. IV. 

It is plain, however, that as far as the form of itself is con- 
cerned, an A.s.f. in -dm may be derived directly from a stem in 
d, and a N.p.m. or f. in -ds from a stem in 4 or d. Nearly every 
supposed contract form may thus be disposed of, if we are willing 
to assume expressly for that form a stem in d or d. 

The stem ndvedas is well authenticated. In i.l65.13<2, eshd'm 
bhdta ndvedd ma rtd'ndm (cf. x.31.3<£, ndvedaso amr'tdndm 
abhUma, and iv.23.4c), ndvedds is N.p.m., and can not be derived 
from ndvedas, except as a contract form. Gr. accordingly sets up 
a stem ndveda for this one single form ; but BR. — as I think 
rightly — refer it to ndvedas. 

vol. x. 73 



550 C. B. Lanman, [as-stems. 

It becomes necessary, then, to examine the usage of the Vedic 
language in order to determine whether we are justified in setting 
up subsidiary stems in d and d, or not. If not, the forms may 
fairly be called contract forms. If the assumption of stems in a 
and d is warrantable, the forms in -ds and -dm may be considered 
as regular derivatives from them. But in some instances, it is 
impossible to decide with certainty. 

I shall now discuss the stems which certainly or probably show 
contract forms, and with them the forms of transition. I shall 
first take up dpd's and jards as typical, and then the rest in 
alphabetical order. 

1. The stem of the Vedic word for 'hope' is dyds, £, from the 
root pas {yans, pds-man) with d'. From it are found in the Rik 
the forms : dpdsd, 4 ; -dsas, N.p., 5 ; -dsas, A.p., 1. There is not 
a trace of the stem dpd'. In the classical language, on the other 
hand, there is not a trace of the Vedic stem dpds; all forms come 
from dpd'. The two stems are undeniably of identical origin ; 
but d-pd' can not be derived directly or phonetically from the 
root pas by any method known to me. How then are the two 
stems dpds and dpd' to be connected ? unquestionably by the 
form dpd'm. This occurs: AV. vi.119.3 : xix.4.2 : Ait. Br. iii.46 : 
vii.26,30: Cat.Br. iii.3.1 16 : vi.7.4': xi.1.6" 8 : xii.4.3"; it may be 
considered in older passages as a contract form for dpasam. 
Corresponding to this A.s. dpd'm, as if the stem were dpd'- (so 
Cat.Br. xiv.9.4"), was formed the transition-nom. s. dpd', ii.1.3 4 , 
4 9 ; 3.1"'": xi.V.l 2 : xiv.4.1 38 ; 5.4 a ; 7.3 3 . In like manner are 
made the forms of the later declensional system: e. g. dpayd, 
Ram. ii.65.35 ; dpe, voc, Bhartr. iii.6 ; dpds, N.p., Cat.Br. i.2.1 12 ; 
dpdbhyas, ib. This stem belongs on p. 494, q.v. ; but is given 
here for completeness. 

2. Somewhat similar is the history of jards, m. In the Rik we 
find : jards-am, 2 ; -dsd; -dsas, Ab., 2 ; and the contract accus. s. 
jard'm, i.140.8 : v.41.17 ; not a single form necessarily implying 

a vowel-stem. In like manner the AV. has : jards-am, i.30.3 : 
ii.13.1 ; -dsd, iii.31.1 : viii.2.8 ; -dse, i.30.2 : vi.5.2 ; -dsas, v.30.17 : 
vi.122.1,4 ; -dsi, ii.10.5 : xii.3.6 ; and the Cat.Br., -dsd, xiii.8.3 1 ' 4 ; 
-dsas, x.4.3 1 . The ace. jard'm occurs AV. viii.2.11 : xix.24.5 : 
TS. iv.3.11 6 : Cat.Br. xiv.6.4 1 . Corresponding to jard'm, as if the 
stem were jard', is formed the transition-nom. jard', AV. xi.8.19 : 
xviii.4.50 : VS. xviii.3 : Cat.Br. v.4.1 1 ; the I.s. jardyd, Cat.Br. 
xii.4.1 1 : xiv.7.1 41 ; the D.s. jard'yai, AV. iii.11.7 : Cat.Br. xi.8.3". 

3. The stem dngiras is of frequent occurrence. In i.83.4, d'd 
dngirdh prathamdrh dadhire vdyah, the N.p.m. dngirds is best 
explained as a contract form. Gr. sets up a stem dngira; but 
the only other Vedic form that occurs to support it is found at 
iv.51.4, where dngire is used instead of dngirasi to preserve the 
formal parallelism (p. 548). I therefore prefer to call drigire a 
transition-form, and not to assume that a stem dngira ever had 
any real existence. For angiro, see A.p.m. 



Transitions. I.] Noun- Inflection in the Veda. 551 

4. Equally doubtful is the stem dn-dga; dn-dgas is Vedic and 
post-Vedic. BR. take dndgds as A.p.m. in vii.60(576).l (so Gr. 
in Ueb.) and 66(582). 4 (so. nah), i. e. as a contract form for 
dudgasas. If this be admitted, why not also dndgds (as N.p.m. 
in vii.87.7 ; 97.2: x.36.12), and dndgdm (gd'm, viii.90.15) ? 
There then remains but one form, dndgdn (iii.54.19: iv.12.4: 
x.12.8) to be explained by false analogy. In x.35.3 (where b, c, 
and d are of 12 syllables), a reads, dyd'vd no adyd prthivi' 
dndgasas; but in 12.8 (where a, b, and d are of 11), c reads 3 
mitrd no dtra dditir dndgdn; so iv.12.4c. Cf. iii.54.195 with 
i.l23.3d 

5. If the text had anehd'[s] in x.61.12, we might take it as 
N.p.m. for anehdsas with kdrdvas; but see N.s.m. 

6. Stem apsards. The stem psdrae is supported by 5 occur- 
rences in the Rik, and by the compounds devd-psaras (3), su- 
psdrastama, and mddhu-psaras ; the stem psara, if at all (p. 
546), only by sapsard'sas, i.168.9, of uncertain meaning. From 
a-psards occur in the Rik : -d's, N.s. ; -dsas, N.p. ; -dsas, A.p., 2 ; 
-dsdm; and in the AV., -dsau, vi. 11 8. 1,3 ; -dsas, iv.37.12 : vi. 
111.4: vii.109.3: viii.5.13: ix.7.10: xii.l. 23,50: xiv.2.9,34; -asas, 
iv.37.3,4; -dsas, iv.3 7.2 : xi.9.15 : xiv.2.35 ; -dsdm, vi.130.1 ; cf. 
gandharvdpsardsas, viii.10.27 bis : x.9.9: xi.7.27; 9.16: xix.54.4; 
-dsas, viii.8.15; "10.27: xi.6.4 ; 9.24; -dsdm, xix.36.6. The A.s. 
apsard'm (AV. iv.38.l5is, 2bis, 4) I regard as a contract form, 
and as the point of departure for the transition-forms apsard'- 
bhyas, ii.2.5 : vii.109.2; -rd'sn, ii.2.3 ; -rdndm, Ram. i.45. 34 (-rasas 
in preceding verse) ; cf. apsard-pat'es, AV. iv.37.7. 

7. The A.s.m. xipdndm, x.40.7 : AV. iv.29.6, is perhaps best 
explained as standing for updnas-am; c£ upanas-e kdvydya, 
Kauc. 139, and BR. s.v. To assume contraction in the L.s. 
(updne, i.51.11, for updnas-i) seems too venturesome. For updnd, 
see N.s.m. 

8. Forms from the stem ushds occur in the Rik by hundreds ; 
but not a single one necessitating a stem ushd'. The A.p. ushd's 
(ix.41.5) and the A.s. ushd'm (i.181.9: x.68.9) are contract forms; 
likewise tcshd'm-ushdm, AV. xii.2.45. Upon such as these are 
based the transition-forms ush'e (du., VS. xxi.17,35 : xxviii.6), 
ushd'bhydm (xxi.50 : TBr. ii.6.14"), and navoshayd (Cak. 175= 
202 Pischel). 

9. Possibly tavd is a contract form for tavds-d (cf. no. 12) ; it 
occurs iv.18.10, tavd-gd'-m vrshabhdm. 

10. Stem ndvedas : see p. 549. For i.79.1, see N.p.f. 

11. In x.92.14, Roth takes nrmdnds as a contract G.s.m. with 
aktds. ' We praise the inviolable Aditi with all her queans and 
the youthful lord of the kindly Night.' Gr., Ueb. ii.5 18, reads 
nrmdndm, A.s.m. 

12. BR. refer puruddnsd, du., vii.73.1, to puruddnsas (q. v. ; 
cf. ddnsas, su-). Whether it is an admissible contraction for 
puruddnsasd is not perfectly certain (cf. tavd, no. 9) ; but if we 
allow it, then puruddnsam, hi 1,23, is a transition-form. 



552 0. R. Lanman, [as-stems. 

13. If mahds be considered as an established adjective stem (so 
with drnas, i.3.12, etc.), then the vexed mahd'm may be explained 
as contracted from mahds-arn; but cf. p. 532. 

14. The forms given by Gr. s.v. medhd' seem to indicate that 
this was originally a vocalic stem. The compounds, however, 
point to a different conclusion. In x.65.10 (where pddas a, b, 
and d are jagatt), c reads br'haspdtim vrtrakhdddm sumedhdsam; 
on the contrary, in 47.6 (where b, c, and d are trishtubh), a reads 
prd saptdgum rtddhttim sumedhd'm \ br'haspdtim matir dchd 
jigdti ; and so x.l25.5<fc=AV. iv.30.3c?, tdm brahmd' nam tarn 
r'shim tdm sumedhd'm. This is certainly a contract form, and 
interesting as called forth by the metre. In vii.91.3a, sumedhd's 
is perhaps A.p.m. 

The Btem mmedhds is well accredited. Besides its occurrences as given by G-r., 
we find: -d's, N.s., AT. v.11.11 : vii.60.1: ix.1.6: YS.iii.41; -as, T.s., Qat.Br. 
xiv.9.1 6 ; -asd, I.S., KatMs. 27.149 and 103.144; -asas, N.p., AT. vii.61.1,2. BR., 
v.900, cite the compounds a-medhas, alpa- {-medhas-drn, Bhag. 7.23), dur- (-medhds, 
18.35), puru-, manda-, (dele sa-,) satya-, hari-. Durmedha occurs as well as dur- 
medhas, but the latter is considered by the grammarians to be the only correct 
form; cf. P&n. v.4.122. Pur-u-medhit-s occurs RT. ix.97.52 ; but ST., i.541 : 
ii.454, has -ds. SwnMha-s occurs at x.132.7, but, as BR. admit, only in a play 
of words with Nr-medha-s; so Nr-medha-m, 80.3. Gr. takes mita-medha from 
medh&. 

In view of these forms, it is fair to regard medhd' m as a con- 
tract form for medhds-am. It occurs in It V. 8 times, in A V. at 
x.2.17, and in SV. i.101 (RV. vedhd'm). Medhd's, N.p., may 
also be a contract form. The remaining three, -d', -dyd (3), and 
-d'bhis, may then be explained as forms of transition. Cf. Benfey, 
SV. Gl. s.v. medhds. 

15. In VS. viii.10, we need not, with Benfey, consider reto-dhd'm 
as a contract form for -dhdsam. See IV. 3. 

16. Stem vayds, ' strength-giving.' I do not think Grassmann's 
article Ivayd! ought to stand. Both BR. and A. Ludwig take 
vayd's in x.92(918).3 as ' Aeste ;' for vayd's, vii.40(556).5, Gr. 
reads in Ueb. i.583, vaydma; there remains only vayd'm, i.165.15 ; 
166.15 ; 167.11 ; 168.10=VS. xxxiv.48. Mahldhara makes it 
stand for vaydsdm. BR. refer it to a stem vayd', ' Starkung,'= 
vdyas (accent !). 

I would assume an oxytone adjective stem vayds. As dpas, 
'work,' is to apds, 'working, busy,' so is vdyas, 'strength,' to 
vayds, ' strengthening, strength-giving ;' cf. Lindner, p. 50. We 
may then take vayd'm as a contract As.m. adj. with ishd-m (for 
vayds-am), and read : 

eshd' ydstshta tanvhe ; vayd'm [id] 

vidyd'mesham, vrjanam jiraddnum. 
' Hither come ye with refreshment for our bodies ; right strength- 
giving refreshment may we receive, (and) well-watered meads.' 
For id, cf. i.l32.5<#, end. 

17. Stem vedhds; see p. 549. 

18. The stem sajdshas is amply supported ; cf. also mjoshas-am. 
I doubt whether the assumption of a stem saj6sha is justifiable. 



Transitions. I.] Noun- Inflection in the Veda. 553 

In iii.4(238).8 and vii.34(550).23, the text has (not -d, but) -ds, 
N.s.f. of stem -as. The N.p.m. in -ds (19) may be regarded as a 
contract form; so -ds in AV. iii.22.1 (or N.s.f.); and djoshds 
(giras, N.p.f.), RV. i.9.4. There then remains only the form 
sajdshau (iii.62.2), which may be explained as a transition-form 
based on the N.p. sajdshds. In VS. xii.74, the text has sajd- 
shasau (cf. BR.) ; but in TS. v.6.4 1 , the transition-form sajdshau. 
As for sajdshds, with apvind in Lll8.ll, and rddasi in iv.56.4, the 
difficulty is not lessened by referring it to sajdsha. 

19. For su-medhds, see no. 14. 

20. A certain instance of a contract N.p.f. is seen in iii.33. 12, 
surd'dhds (nadias), for surd'dhas-as; cf. BR. 

Benf'ey, Abh. xvii.22, would allow contract forms in the classical 
language. He cites : ati-yapdm, A.s. fern., for ali-yapas-am, 
Nala viii.4; cushka-srotdm (nadim), for -asam, xvi.14. As tran- 
sition-forms might then be regarded: kdrta-yaca-m, Arsh. Br. 
i. 36,265,545 ; prati-srota-m, Bhag. P. x.78.18; cf. sahasra-srota-, 
ib. v.20.26. 

II. Further Transitions from the 4S-Declension. 

Besides the regular as-forms, there occur a number of shorter 
equivalent a-forms, but without any contract forms to mediate 
between them. Some of them look like forms of transition whose 
point of departure was the N.s. 

1. The ambiguity of the N.s. is illustrated by rdkas, vi.66.6. 
It is impossible to tell whether it is N.s.n. of a stem rdkas, or 
N.s.m. of a stem rdka. 

2. Hedas is generally neuter. In vi.62.8, it is marked as neuter 
by ydd; and so in AV. xix.3.4 by the construction (A.s.n.) ; but 
in VS. xiii.45, it can be regarded as N.s.n., or as N.s.m. {heda-s). 
In RV. i.94.12, on the other hand, it is marked as masculine by 
ddbhutas. This is the point of contact for the two declensions. 
Based upon such a form as masc. heda-s we find : h'eda-m, AV. 
xii. 4.20,21 ; hide, L.s., RV. vii.62.4. 

3. An excellent illustration of the coincidence of certain forms 
of the as and a-declensions is seen in the stems cdrdhas and 
cdrdha. Gr. assigns 33 Rik-forms to the stem cdrdhas, and 26 
to cdrdha; but it is impossible to decide in some cases. Thus 
Mtlller, Translation p. 57, takes cdrdhas as N.s. neuter in ii.1.5, 
tudm nard'm cdrdho asi purtivdsuh; but Gr., as N.s. masculine. 
The same uncertainty exists in iii.32(266).4. So in v.33.5, ndrah 
cdrdho jajMnd'h, BR. take cdrdho as N.s.m. ; but Gr., as N.s.ii. 
So in v.54.6, dbhrdji cdrdho, maruto, ydd arnasdm \ mdshatha: 
Gr. refers pdrdho to cdrdhas in Wb. ; but' in Ueb. he makes 
arnasdm object of mdshatha, so that cdrdho can be referred to 
cdrdha. Cf. also i.37.1 with v.56.9. These collocations seem to 
me to account for the series of a-forms, cdrdha-m, cdrdhena, -dya, 
-asya, -dn. See also Mtlller, Translation pp. 56-8. Cf. below, 
no. 8. 



554 C. R. Lanman, [as-stems. 

4. The isolated a-form dvena (i.128.5) may be based upon the 
common nom. dvas, since this might sometimes be felt as a masc. 
(dva-s, v.70. 1). Cf. nir-avdsya, i.122.11. 

5. In the Rik we find the neuter stem krdndas. In the AV., 
the N.s. krdndas may be taken as a masc. in xi.2.22 {krdndas). 
Based thereupon is the D.s. krdnddya, xi.2.3. So p&ci-kranda-m, 
RV. vii.97.5. In AV. xi.4.2, the metre favors the restoration of 
the old as-i orm : ndmas te prdna krdndds-e, text krdnddya. 

6. Forms from the stem saprdthas are common. Now in TBr. 
i.1.10 2 "*, we find a series of 5 formulae : dtharva pittim me gopdya/ 
ndrya prajd'm m. g.; pdnsya papil'n m. g.; sdpratha sabhd'm m. 
g.; dhe budhniya mdntram me gopdya. One needs no argument 
to be convinced that the sporadic a-form is due simply to formal 
parallelism (cf. BR.) ; and I may here add that the instructively 
mutilated form dtharva is due solely to the same cause. The VS., 
iii.37, has dtharya ; for this, the Brah. stupidly intends to sub- 
stitute the more intelligible dtharvan, whose correct vocative 
could only be dtharvan. 

7. Forms of the stems pepas and su-pepas are common. Pepas 
is N.s.n. of pepas in iv.36.7 (preshtham), and A.s.n. in i.6.3 (cf. 
a-pepds-e) ; but in vii.34.11 and 42.1, it may be N.s.masc. (pepa-s). 
In Ait. Br. iii.10, we find : pepas (A.s.n., 3) and pepasd; but also 
pepds once as N.p.m. For puru-pepdsu, see p. 548. Cf. also 
paSpd yd (N.s.m.) and ptehr-patpe-m {nmdne-m, N.s.n.), and Justi, 
Handbuch p. 1816. 

8. The influence of the requirements of the cadence and of 
formal parallelism is plain in the use of pdrdhd- and vdrpd-, 
rather than pdrdho- and vdrpo-, in iii.34.3 : 

indro vrtr&m, avrnoc ch&rdhanttih 
pro, mdyindm arnindd varpanitih. 

9. The singular form card'yai is called out by the requirements 
of the cadence. Cf. 

v.47.4, ddfa g&rbham car&se dhdpayante, 
i.92.9, vipvam jivam car&se bodh&yantt, and 
vii.71.1, vif vam jtvarii pr&mv&ntt card'yai. 
The short a of prds- excludes it from the eighth place and makes 
the order cardse prdsuvdnti metrically impossible. 

10. Compare tdras-e, iii.18.3, with tdrdya, ii.13.12. 

11. The neuter dveshas is very common; for a-dveshd, seep. 
548. 12. Besides forms from jrdyas and prthu-jrdyas, we find 
once prthu-jrdya-m rdtham, iv.44.1. 13. So besides 6kas, ni- 
okas, sdm-okas, we find dur-oka-m, vii.4.3. 14. With drddhas4, 
TS. iii.2.2 3 , cf. vi-dradh'e, du. f., RV. iv.32.23. 15. With H'rna- 
mradas-am, VS. ii.2, cf. Urna-mrada-m, Kau§. 137. 16. With 
dvibdrhas, cf. dvibdrha-jman. 

The locatives pi. in -asu for -as-su are perhaps made after the 
analogy of the aw-stems. 

III. Further Transitions trom the 4&-Declbnsion. 
The transition-stem (in asa) is longer than the original stem (in 
as). The transition in Pali results sometimes in a shorter, but 



Transitions. III.] Noun- Inflection in the Veda. 555 

sometimes also in a longer form ; thus from mahd-\-anas we have 
mahdna-m and mahdnasa-m. The same is true in Sanskrit. 

1. In Katy. Or. xiv.2.31, we find mahdnasdndm . . . mahd- 
nasa-s mahdrathah. 2. Corresponding to the N\p. tigmd- 

tejas-as (AV. xix.9.10), occurs the A.s. neuter tigma-tejasa-m 
(astram dtishtham, Arj. vii.20). 3. A like transition-form is 
su-varcasa-m (patarh sutdndm) ; see BR. 4. Upon the adverbial 
compound d-jaras-dm, with adv. accent, rests the form djarasd'ya 
(RV. x.85.43, accent). 5. BR., iv.921end, cite prd-cetasdndm 
as occurring in the Brahma P. for prd-cetas-dm; but LA. (Bonn, 
1865, p. 58, 1. 8) reads prdcetasdndm. The accent of arnasd, 
avasd, and parinasd seems to mark them as secondary derivatives 
of drnas, dvas, and pdrinas. 

In Yasht x.143, we find the normal form (vdshem, masc.) ptehr- 
paSpanh-em, i. e. str'-pepas J am. This comes to be. used as a 
neuter also ; and we have, e. g., vanhanem ptehr-paSpahhe-m, 
i. e. vdsanam str'-pepasa-m, xiii.3 ; and likewise aiwydonhanem 
(i. e. avya/igam) ptehr-paSpanhe-m, Yac. ix.26. For vippem paiti 
ushdonhe-m (?), masc. !, see p. 546. 

IV. Transitions or Radical Stems to the .AS-Declension. 

The N.s.m. in -ds has been the point of departure for the tran- 
sition of radical stems to the as-declensiou. See p. 549. Here 
belong : 

a. the stem candrd-mds (no. 1) ; 

b. compounds of the verbals -dhd' and -dd' (2-6) ; 

c. compounds of radical substantives (7-11). 

1. For good practical reasons, the dictionaries set up a stem 
candrdmds. The true stem is candrd-mds. Of this only the 
N.s. is regular — candrdmds (RV., 5 times, and AV., 19). The 
other cases ought also to have the long d throughout, candrd- 
mdsi, etc., just as well as sd'rydmd'sd. In fact we find : candrd- 
mds-am, Cat.Br. vi.7.4 T ; -mdsd, AV. xiii.4.28 ; -mdsas, RV. 
i.84.15; -mdsi, AV. x.3.18: xi.5.13; candramds, AV. vii.81.2; 
candrd-mdsd, RV. x.64.3 ; stirydcandramd'sd, i.102.2; -aw, v. 
51.15: x.190.3. Seep. 547. 

2. It seems to be adverse to all Vedic analogies to assume a 
stem like vayo-dhd's; but Benfey assumes such, and even refers 
vayodhd'm to a stem vayodhd's (for vayo-dhd 's-am, SV. Gl. — cf. 
Gram. p. 316, 1. 16). There are nigh 200 stems like vayo-dhd'; 
see p. 438. Of this, the N.s.m. is vayo-dhd'-s (cf. p. 443). To 
this now as a point of departure I refer the following forms of 
transition : vayo-dhd'sam (indram), VS. xxviii.24 ; -dhdsd, I.s., 
VS. xv.7; -dhase (indrdya), Vaitana Sutra xxx.25 : Katy. Or. 
xix.5.22 ; -dhdsas, 1ST. p.m., AV. viii.1.19 ; -dhas (soma), RV. 
ix.81.3. Vayodh&s can not be a direct voc. of -dhd' (cf. BR.), for 
this would be vayodhds (p. 449). 

3. I think no Vedic form necessitates the assumption of a stem 
reto-dhd's. The Rik has 5 forms from reto-dhd'. So in TBr. ii.7.4 1 , 



556 0. R. Lanman, [as-stems. 

we have retodhd-s somas and retodhd-s eshd. The only form that 
seems to require an as-stem is reto-dhd' s-as, G.s.m., VS. viii.10 ; 
but this is best ascribed to false analogy. 

4. The A.s.f. varco-dhd'-m occurs VS. iv.ll ; the N.s.m. varco- 
dhd'-s, AV. ii.11.4. The D.s.m. varco-dhd's-e, AV. iii.21.5, is a 
transition-form. 

5. Forms from puro-dhds occur : -dhasdm, Bhag. x.24 ; -dhasd, 
Oak. 71.18 (Bohtl.) ; but they are only post-Vedic. 

6. The Rik has 30 forms from dravino-dd' . The N.s.m. is 
-dd'-s, and corresponding to this, as if the stem were -dd's, are 
made the transition-forms -dase (devdya), Schol. to Katy. Cr. 
ix.13.19 ; -dds, V.s., RV., 6. For "drdvinodasds" see Ab.s.m. 

7-9. It must be admitted, I think, that the masculine compound 
of a fern, substantive in d may make its N.s. in -ds (see p. 439 ad 
init.) ; otherwise, we should have to assume a stem puru-nishthti! s 
for the N.s.m. -nishthd's. On p. 439 I referred the forms bahu- 
prajd'-s and suprajd'-s (N.s.m.) to stems in d, BR. refer them to 
-prajd's; cf. also Pan. v.4.122,123. I would further refer to cisterns 
the following forms, which the reader is requested to add on pages 
439, 443, 444: N.s.m., dprajds (apaptis), Cat.Br. i.6.1"; suprajd's, 
AV. iv.11.3: VS. iii.37: vii.18; divdkshds, RV. iii.30.21 ; N.s.f., 
suprajd's, AV. xiv.2.22,23,74. 

As forms of transition I regard: (7) dprajasam, A.s., AV. 
vii.35.3: xii.5.45 ; aprajasi, L.s.f., Yajn. ii.144; (8) ida-prajasas, 
A.p.f., TS. i.5.6 1 ; (9) suprajdsam, A..s.f., AV. xiv.1.49; -dsas, 
N.p.m., AV. iii.10.5 : xviii.4.63 : TS. Ll.10* : iii.2.4 6 . 

For stem dushprajas I find no example ; stem dushprajd occurs 
Bhag. P. x.49.4. The compound suprajdstvd does not prove the 
existence of a true stem suprajd's any more than suprajdstvd 
proves a stem suprajd's. Cf. also aprajdstd and aprajasyd with 
aprajdstvd and andgdstvd. 

10. The forms cited by BR. and Gr. in the article divd-kshas I 
treat otherwise. On account of the accent^ I set up a bahuvrthi 
stem divd-kshd, ' having a dwelling (kshd 1 ) in heaven ;' cf. div'i- 
yoni. Of this, divd-kshd-s, iii.30.21, is the N.s.m. (cf. pp. 439, 443) ; 
and the N.p. divd-kshasas (m., x.65.7 ; £, iii.7.2) may be regarded 
as due to false analogy. 

11. The A.s.f. kshd'm (p. 446), in i.67.5 ; 174.7: vi.6.4: x.31.9, 
and the A.p.f. kshd's (p. 452), in iv.28.5, have the value of two 
syllables. Gr. reads kshd'mam, kshdmas. Perhaps the true 
resolution is kshds-am and kshds-as (transition-forms based on the 
N.s.f. kshd'-s), for which the diaskeuasts have written the contract 
forms kshd'm and kshd's, coinciding with the regular forms of 
kshd'. 

Y. Transitions of ^4-Stbms to the .AS-Declension. 

Under section II. we saw neuters like h'edas passing into the 
a-declension (h6da-m). We have now to discuss the opposite 
movement and to group together a number of well authenticated 
a-stems, by the side of which occur infrequently or sporadically 
as-forms. 



Transitions. V.] Noun- Inflection in the Veda. 557 

1. Thus besides pepa (RV. — see BR.), the AV. has also cepas. 
Perhaps the* N.s.'m. pepa-s was regarded as a neuter form, and the 
A.s. made accordingly. The Rik, for example, reads ydsydm 
updntah prahdrdma pepam, x.85.3ld; while the AV. has prahd- 
rema pepas, xiv.2.38<?. 

In some of the following cases, however, I am not sure that the 
relation is strictly one of transition ; and perhaps the metre has 
had more to do with calling forth the secondary forms here than 
elsewhere. 

2. Forms from vepd are common. The only support for vepds 
is in AV. v.23.12, hatd'so asya vep&'s-as \ hatd'sah pdrivep&s-as, 
where vepd'sas would not fit the cadence. 

3. Instead of the regular -posha-m, we have -poshas-am, for no 
other reason than to fill out the verse; see page 548. 

4. A like instance is svasti adyd ushdso doshdsap ca, AV. 
xvi.4.6 ; cf. p. 468. 

5. The stem ddksha is abundantly authenticated ; so samdnd- 
daJcsha and p-Atd-daksha. On the other hand, ddkshas-e occurs 
thrice ; pUtd-dakshas-d and -as, 1 times ; and ddkshas-as at viii. 
13.1, for which the SV, at i.381 and ii.96, has ddkshasya. 

6. Forms from bhdra are common ; but twice we find stem 
sd-bharas, and once vipvd-bharas-am. 

7. For svand and compounds, see BR. Twice we find tuvi- 
shvands, and once vd'ta-svanas (? see p. 546). 

8. So besides the stem svard and its compounds (see BR.), we 
find gharmd-svaras-as, A.p.f., iv.55.6. 

9. The stem c6ka and its compounds (see Gr.) are well authen- 
ticated ; but once we have sahdsra-pokdn, N.s.m. 

10. For peva and its compounds, see Gr. RV. x.85.246 reads 
y'ena tvd'badhndt savitd! supevdh; but the Atharvan variant 
(xiv.l 19,58), . . . sup'evdh; so ii.2.2. 

11. The form dhruvds-e in vii.70.1, a" ydt seddthur dhruvdse nd 
ydnim, is best explained as a simple transition-form from dhruvd 
(cf. BR. iii.1002, 1004), preferred to dhruvd'ya on account of the 
metre, and also because a form in -dse is common for datives with 
infinitival function. Cf. Hariv. 3959, tarn giripreshtharh sve 
sihdne nivepaydrndsa dhruvdya. 

12. With d'evd vrdhd'ya hUmaJie, viii.72.6, cf. (stotfndam ca 
spttrdhdse, v.64.4e?, and) sdkhinaam ca vrdhdse, v.64.5<£ 

13. With ddha (common as post-Vedic), cf. the an. X. ddhas-d 
in the same signification, x.11.1. 

14-16. With rdva, vrshd-ravd, cf. tuvi-rdvas, puvd-rdvas. With 
toed, cf. topds (? p. 545). With sahd, cf. sdhas, as adj., BR. 
vii.866 end. 

VI. Miscellaneous Cases of Transition. 

1. In i. 141.8^, c-d'rasyeva tveshdthdd ish&te vdyah, the ninth 
place requires a short syllable and so ishante is excluded. The 
lack of concord between the singular verb and its plural subject 
vdyas, 'fowls' (from vi), was the less noticeable inasmuch as 

vol. x. 74 



558 C. R. Lawman, [as-stems. 

vdyas itself was felt as a collective noun, ' fowl, Gefltigel.' In 
this case vdyas would easily come to be regarded as a singular 
form (N.s.n.) of a new transition-stem vdyas. So in i.104.1, vdyas 
may be taken as A.p.m. (p. 395 med.), or as collective A.s.n., 'the 
flying team' of Indra. Such passages seem to explain the origin 
of the stem vdyas/ but it is not absolutely necessary to posit this 
stem, except for texts later than the Rik. 

2. If, in v.37.15 {djdhvdno ghrtdprshthah sudncdh), we had 
sudnctirh, this might easily be referred as transition-form to 
sudnc-arn (agmm, vi.15.10), i. e. su-dnca-m. As the text stands, 
perhaps we have a an. \. of independent formation ; but cf. the 
variants su^'evdh and su$v&h, above, V.10. 

3. Besides the usual stem rddasi, we find rSdas once, ix.22.5a, 
eti. prshthd'rii rddds-os (-- - ^ -) ; it is used simply and solely 
because rddasyos would not fit the cadence. With good reason 
Gr. reads rddasos (for -syos) in the cadence of i.l51.3a and 168.1c. 

4. So instead of the usual sdrasa, we find stem sdras once, 
MBh. xiii.736, krkaldsakdsdrdsdm, — used probably for the sake 
of the cadence. Otherwise, it may be taken as a shortened 
G.p.m. of sdrasa, for sdrasdndm (see p. 353). 

5. The relation of aJcshi-6s to aksh-6s is like that of r6dast-os to 
r6das-os ; but in AV. v. 4. 10, where aJcshrds occurs, the metre 
demands ahshi-6s — see p. 392. 

Compare md'ns and vandMir, p. 494. It is hard to see what 
relation, if any, exists between sarasi' and sdras. 

Nominative Singular Masculine and Feminine. 

I. Here belong 341 masculines (from 110 stems), and 86 femin- 
ines (from 24 stems). 

Masculines : angwds, 5 ; acetd's ; ddribarhds ; advayds ; andgds (437, 2 ; 602 
7) ; anudhd's ; anunavarcds ; anend's, 3 ; apracetds, 3 ; dbhibh&tiojds, 5 ; dmitau- 
jds ; arapd's, 3 ; dvaydtdhedds ; dsamdtiojds ; dhand's ; urucakshds, 2 ; uruvydcds, 
3 ; r'shvmands ; rshva&jds ; kdrudhdyds, 4 ; Jcelavedds ; Jthd'das-arnds ; gabhtrd- 
vepds; gurtdmands ; g-tbrtacravds ; gonioghds; ghrtdprayds; jdtavedds, 23; tddapds, 
2; t&dokds; tadojds; tardddveshds ; trivayds ; ddbhrdcetds, 2; damunds, 18 
dasmdvarcds, 3 ; dirghdtamds, 2 ; dirghd'psds ; durosMs; devapsards; devdcravds 
ndvedds (319, 4; 366, 3); nrcdhshds, 18; nrmdnds, 4; nodh&'s, 3; niokds: pdva- 
kavarcds; prthujrdyds ; prthupd'jds, 3; prdcetds, 10; prdvayds ; bdhuejds; brhdc- 
chravds ; bodhinmands ; bhu'riqjds ; mddhuvacds, 2; mdnqjavds, 6; mitrdmahds; 
yacd's, 2; rakshd's ; raghupdtmajanhds ; rifd'dds, 2; vdsufravds; vd'taranhds, 
2; vicetds, 7; vimands, 2; vicvd-cakshds, 2; -dhdyds, 2; -bhqjds; -mands; -vedds, 
9 ; mcvavjds ; vihdyds, 7 ; vidudvishds ; vtduhards t ; vrddhdrmahds, 2 ; -vayds ; 
-cravds ; vedhd's, 14 ; cukrdvarcds ; craddhd'mands ; sdcands ; sacetds ; sajdshds, 
26; satydrddhds, 5; saprathds, 4; sdmokds, 2; sahasra-cakshds ; -cetds; -pdthds , 
-retds, 3; -fokds; sahasrdpsds; suddnsds, 3; sumdnds, 25; sumedhd's, 11; surd'- 
dhds, 4 ; sureknds ; suretds ; suvdrcds, 3 ; suvd'sds ; stomavdhds ; spdrharddhds . 
suancds; sudpds,S; svdbhutiojds ; svdyacds,3; suarcakshds; suarccmds; sudjds,2 

Feminines: Advayds; apsard's ; arepd's, 3; dhand's; uruvydcds; usM's, 50 
u'rnamradds ; nicd'vayds ; nrmands, 2 ; prdcetds ; mddhuvacds ; ydvayaddveshds 
vihdyds; vr'shamands ; cukr&vdsds; sacetds; sajdshds, 4 (i. e. 490, 7 ; 395, 8 
238, 8; 550, 23) ; saprdthds ; sabhards ; suddnsds; supScds, 2; sumdnds, 2 
suvd'sds, 6 ; sumedhd's, lii.5'7.5 (with jihvd', ' voice,' BR.). 



N.s.m.f.] Noun- Inflection in the Veda. 559 

Thrice Gr. proposes to resolve -ds into two syllables. In no 
case is it necessary. Prdcet&s (Gr. -aas) is at the end of a cata- 
lectic pdda of 7 syllables in vi.14.2a, and of 11 in vii.16.5c. In 
iv.21.6c, read 61 durdshdh paastidsya hdtd. 

Candrd-mds (RV., 5; AV., 19) belongs properly on p. 495; see 
p. 547; divd-leshds and suprajd's (m., AV. iv.11.3), on p. 443; 
suprajd's (f., xiv.2.22,23,74), on p. 444. For ■d 'rnamradds, g-drtd- 
vacds, dvibdrhds, and vtrdpepds, see N.A.s.u. For dndgds (076, 
1 and 582, 4) see p. 551. In i.79.1, ndvedds is N.p.f. For 
prdcetds, vi.5.5, see V.s.m. 

Transition-forms (see pp. 550-557) : I. dpd'; jard'; updndf; 
medhd'; V. sdbhards; sahdsrapokds ; supevds. 

II. The instances in which the s of the N.s. is lost are doubtful. 
See Pan. vii.1.94. The least doubtful is Updnd (RV., 8; TS. 
ii.5.8 5 ) ; but is not this a transition-form, after the analogy of 
feminines (cf. JVdmi, Pr'tht, etc., p. 375 med.), based on the con- 
tract ace. updndm (for -asam) ? Of like sort, possibly, is (kdrdvo) 
anehd', x.61.12. Of purudanpd I find no occurrence. The form 
r'bhvd need not be referred to the same stem with r'bhvas-am. 
For ayd', i.87.4, see p. 358 end. 

III. Several peculiar forms claim our attention. The samhitd 
has: tuvirdvdn k-, x.64.4; -dn r-, 16; — svdtavdn r-, iv.20.6 ; 
svdtavdnh p- (Prat, iv.34), iv.2.6 ; svdtavdnp ca, VS. xvii.85 ; — 
svdvdn, before a-, vi.47.12: x.131.6 (AV. vii.91.1) ; i-, vi.47.13 : 
x.131.7 (AV. vii.92.1) ; r-, iii.54.12 : vi.68.5 ; e-, x.92.9 ; y- (Prat. 
iv.28), i.118.1; 35.10. The VS. variant (xxxiv.26) of the last 
passage has svdvd ydtu (Pr. iii.135). Similarly Roth would read 
sut[d]vd ydd, x.99.11 ; see p. 488 end. 

The treatment of the forms in sandhi is precisely like that of 
the forms on pages 512, 514, and 517. The pada has in every 
instance tuvfardvdn, svd-tavdn, svd-vdn. 

The stem of the second word is svd-tav-as, as is abundantly 
shown by the other forms ; that of the third, sti-av-as, as shown 
by the metre and the other forms. BR. refer the first to tuvi- 
rdvant (see p. 517) ; but I think Aufrecht is right in setting up 
the stem tuvi-rdvm, like purd-rdvas, brhdd-ravas, Z. D. M. G. 
xxv.233. 

Aufrecht, ib. xiii.501, considered all themes in as as ultimately 
derived from themes in ant, and explained these nominatives as 
relics of the more ancient themes in ant. (For svdtavadbhyas, 
etc., see I.p.) I think that they are formed after the analogy of 
the stems in vans (vas) and vant, the radical v of the stems 
sn-dv-as etc. being felt as part of the derivative ending (svd-vas). 

The same misconception has worked elsewhere. As a result of 
the misdivision of bhdv-ant into bkd-vant is perhaps to be ex- 
plained the voc. bha-vas (p. 509). In like manner I account for 
the peculiar fern, to mdtarv-bhv-an, viz., mdtari-bh-vari, N.p. -is, 
x. 120.9 ; Say., mdtari bhavantyah. Cf. mdtari-p-varis (root $€), 
AV. xx.107.12; and mdtari-p-vas /, voc. Upon a confusion of 
the final element of anud-vdh with that of the vant-stems rest the 



560 0. R. La?iman, [as- stems. 

peculiar forms cited on p. 499 (anadrvd'hp etc.). Cf. yuvat and 
kshd'm-an, p. 530. The explanations of Weber (Beitrage iii.388) 
and Brugman (Zeitsch. xxiv.7l) are essentially like the above. 

Nominative and Accusative Singular Neuter. 

I. Here belong 1409 forms (from 101 stems). 

Enumeration : drihas, 33 ; anjas, 2 ; adveshds, 4 ; Anas, 8 ; anehds, 5 ; dndhas, 
3; dndhas, 19; dpas, 15; apds; dpnas, 4; dpsas, 3; dmbhas; dyas,i: arakshds; 
arapds, 2 ; drqas, 15 ; dvas, 73 ; d'gas, 17 ; d'pas, 1 ; i/ras, 2 ; A'dhas, 4 ; enas, 
19; dkas, 16; #as, 53; kshddas, 10; cartas, 12; chdndas, 3; jdnhas; juvas ; 
jrdyas, 8 ; tddapas ; tdpas, 4 ; tdmas, 43 ; Mras, 4 ; tf$as, 2 ; tydjas, 3 ; ddnsas, 
4; duvas, 13; drdvinas ; dveshas, 25; ndbhas, 11; nawias, 41; pdkshds; pdyas, 
52; pd'jas, 11; pd'thas, 14; pi'vas, 5; purubhdjas ; picas, 6; prdthas; pray as, 
21; psdras, 3; bhdrgas, 3; bhd'sas, 2; nwinas, 76; mayas, 19; mahds, 14: 
m<$ftas, 8 ; mr'dhas, 2; raidas; ydfas, 24; rdkshas, 20; rrf/(W, 38; rdpas, 12; 
rdbhas; rd'dhas, 61 ; reknas, 10; refos, 21 ; repas ; rodhas, 2 ; vdkshas, 4; i;aca«, 
47 ; vdyas, 2 ; vdyas, 70 ; vdyas-vayas, 2 ; vdrivas, 26 ; vdrcas, 6 ; vdrpas, 7 ; 
vd'sas, 8; vd'has, 3; flate, 14; vepas ; vydcas, 3; vrdyas ; cdrdhas, 28; panas, 
34 ; pir&s, 25 ; ceshas, 6 ; provas, 87 ; sajdshas ; sodas, 13 ; sddas-sadas, 2 ; sa- 
prdthas, 17 ; sabd'dhas, 14 ; saras, 4; sdhas, 48; swStas; srdias; svdtavas; haras; 
haras ; hedas, 6 ; hvdras, 2. 

Transition-forms (see pp. 554-558) : II. durdkam; III. tigma- 
tejasam; suvarcasam; V. pepas, A.s. ; VI. vdyas. 

II. In 20 or more instances, the form in <2s is used rather 
anomalously as a neuter. Gr. admits only one such, devdvyacds, 
iii.4.4. Here Benfey (iJA. xix.260 ; cf. 255 ff.) considers that 
the original form was the regular one in -<fcsy that the final 
visarga disappeared tracelessly ; that <fc was then lengthened 
metrically in the samhitd; and that to the misapprehension of 
the forms in -<2 on the part of the padakdra are due the^?a<Za- 
readings in -<2A. But this explanation does not fit the other 
cases well. 

Here belong : devdvyacds (barhis, cf. i.142.5), iii.4.4; dvibdrhds, 
(vdyas) i.71.6 ; (pdrma) i.l 14.10 ; (sd'rna) iv.5.3 ; (vdoas) vii.8.6 ; 
as adv. (see BR. v.28), v.80.4 : vi.19.1 : vii.24.2 : x.116.4 ; vird- 
pepds (drdvinam), iv.11.3 : x.80.4 ; gUrtdvacds (brdhma), x.61.1 ; 
viduhards (f tdpas), x.109.1 ; U'rnamradds (barhis), v.5.4 : so 
VS. xxi.33,57: TBr. ii.6.11 8 ,14 6 ; vipvdvyacds (cdrma), AV. ix. 
7.15. In AV. i.26.3, the MSS. have pdrma saprdthdh; ed., -dh. 
Compare also: AV. xviii.2.19, ydchdsmai pdrma saprdthdh; 
T.ar. x.1.10, ydchd nah pdrma saprdthdh; and RV. i.22.15, -dh; 
and finally, pdrma saprdthdh, VS. xviii.54 : TS. iv.7.13 3 . The 
neuters patasd's, suprayd's, sthd's, indra-, soma-jd's, and presh- 
thasthd's, p. 445, should be considered in connection with the 
above forms. 

Accusative Singular Masculine and Feminine. 
Here belong 217 masculines (from 97 stems) and 58 feminines 
(from 22 stems). 

Masculines: acetdsam, 2 ; dnashtavedasam ; dnagasam; dnugdyasam; anehdsam, 
8; apdsam; dpratidhrshtacavasam ; abhibh&Udjaaam ; ardmanasam; arddhdsam, 
2 ; arepdsam ; arcand'nasam ; dhandsam ; updkdcakshasam ; uru-cdkshasam, 2 ; 



A.s.m.i] Noun- Inflection in the Veda. 561 

-jrdyasam, 2 ; -vyacasam, 2 ; ksketrasd'dhasam ; gdthapravasam ; gdyatrdvepasam 
girvanasam, 4 ; girvdhasam, 2 ; gtirtdpravasam ; gavarnasam ; godhdyasam 
cikitvinmanasam ; citrdmahasam ; citrdrddhasam ; jardsam, 2 ; jdtdvedasam, 24 
tavdsam, 13 ; tuvird'dhasam; tuvishvandsam ; tuviqjasam; tyajdsam; dam&nasam, 
5 ; ddnd-okasam ; dyukshdvacasam ; dvibdrhasam, 3 ; dvicavasam ; dhrshnuojasam 
ndridpasam ; nrcdkshasam, 4 ; panktirddhasam ; pdrtnasam ; pv/ru-pepasam 
-bhojasam, 2 : -vdrpasam ; -vepasam ; prdcetasam, 8 ; prdtvakshasam ; brdhma- 
vdhasam ; bhiydsam, 3 ; bh&'ri-cakshasam ; -dh&yasam ; yajnd-vanasam ; -vdh 
asam ; yapdsam, 15 ; rakshdsam ; ripd'dasam, 3 ; vd'japravasam ; vicetasam, 2 
vijoshasam ; vidveshasam ; vipvd-dhdyasam, 4; -bharasam; -vedasam, 6; -vyac 
asarn ; vipvd'yu-poshasam, 2 ; -vepasam ; vihdyasam, ; vttirddhasam ; vedhdsam 
sdcetasam ; sajoshasam ; satyd-girvdhasam ; -pavasam ; sddmamakhasam ; samu- 
d/rd-vdsasam, 2 ; -vyacasam ; sahdsra-cakshasam, 2 ; -bharnasam, 2 ; -varcasam. 

2 ; svxetasam, 2 ; suddiisasam, 2 ; supepasam, 1 ; supraydsam, 3 ; subhojasam 
swnedhasam ; surd'dhasam, 4 ; suretasam ; suprdvasam, 3 ; srprdbhojasam 
svdyapasam, 3 ; sudvasam, 3 ; hdridhdyasam ; hdrivarpasam. 

Feminines : dndgasam ; aneh&sam ; arakshdsam, 2 ; arepdsam. 2 ; dpvopepasam 
udojasam ; uruvydcasam; cikitvinmanasam; dhvardsam; pwubhdjasam; ydvaydd- 
dveshasam ; vd'japepasam ; vipvaddhasam ; vipva-dhdyasam, 2 ; -pepasam ; -bhoj- 
asam ; patdrcasam ; pucipepasam ; sahdsrabharnasam, 2 ; svdyapasam ; hdrivar- 
pasam ; ushdsam, 33 ; ushd'sam, p. ush&'sam (page 546), 11. 

Adverbial shift of accent is seen in djarasdm, Qat.Br. i.6.3* 1 : 
cf. Ait. Br. i.28. See p. 495. 

Contract-forms: vedhd'm, ix.26.3 ; 102.4; updndm, x.40.7: 
AV. iv.29.6 ; mahd'm f, 23 times ; sumedhd'm, x.47.6 ; 125.5 ; 
vayd'm,l 165.15; 166.15; 167.11; 168.10: VS. xxxiv.48 ; dpd'm, 
AV. vi.119.3 ; jard'm, i.140.8 : v.41.17 ; dndgdrn, viii.90.15 ; 
apsard'm, AV. iv.38.1fo's, Ibis, 4 ; ushd'm, i.181.9: x.68.9: AV. 
xii.2.45 ; medhd'm, 8 times, and AV. x.2.17. For vayd'h, vii. 
40.5, see p. 552, no. 16. 

Transition-forms (see pp. 551-557) : I. puruddnsam ; II. h'edam; 
pdrdham ; p&cikrandam ; prthujrdyam ; IV. candrdmdsam ; 
vayodhdsam; dprajasmn; suprajdsam; kshdsamf; V. vi$vd'yu- 
poshasam; vipvdbharasam; tuvishvandsam. 

Instrumental Singular Masculine, Feminine, and Neuter. 

I. There are 36 masculines (from 15 stems) ; 30 feminines (from 

3 stems) ; and 704 neuters (from 66 stems). 

Masculines : anehdsd ; arakshdsd ; gopartnasd ; jardsd ; tdr&shasd ; tavdsd ; 
pdrtnasd, 6; prthupd'jasd, 2; bhiydsd, 11 ; mdnojavasd; yapdsd ; vipvdpepasd ; 
sdhasd, 2; supepasd; suprdvasd. 

Feminines : arepdsd, 2 ; ushdsd, 2 V ; yajdsd. 

Neuters : dnjasd, 4 ; dnasd, 3 ; dndhasd, 13 ; dpasd, 2 ; apdsd ; arakshdsd ; 
drnasd, 2; dvasd, 52; enasd; qjasd, 104 (but cf. G.s.n.) ; dhasd; kshddasd, 3; 
gdvarnasd; cdkshasd, 16; cetasd, 6; jdvasd, 5; javdsd; tdnasd (p. 4T9); tdpasd, 
10; tamasd, 13 ; tdrasd; te)'asd,3; iydjasd, 4; tvdkshasd, 3; tveshasd; ddnsasd, 
2 ; dohasd ; dhd'yasd ; dhrdjasd ; ndbhasd, 2 ; ndmasd, 84 ; pdyasd, 32 ; pd'jasd, 
5 ; t pi'vasd ; prdyasd, 5 ; bdhmjasd ; bhrd'jasd ; mdnasd, 115 ; mdhasd, 3 ; 
medasd; yapdsd, 3; rakshasd; rdjasd, 4; rapasd, 4; rd'dhasd, 16; rffcnasd; 
retasd, 3 ; vdcasd, 10 ; vdnasd ; vdyasd, 6 ; vdrcasd, 5 ; vdrpasd, 4 ; vd'sasd, 2; 
vd'hasd, 2 ; vedasd ; vepasd, 2 ; pdvasd, 80 ; peshasd, 2 ; crdvasd, 9 ; sdhasd, 29 ; 
suddnsasd; surMasd; srdtasd; hdrasd, 6; havdsd, 2; heshasd. 

The masc. bhiyds-d is contracted to bht'sh-d; and this appears 
with adverbial shift of accent in i.l33.6fos <-d'n a-) : vii.21.3 : 
viii.86.14. Cf. p. 494. 



562 0. R. Lanman, [as-stems. 

Contract form: tavdm tavd-gd'mf; see p. 551. 

Transition-forms (see p. 550-557): I. jardyd ; navoshayd ; 
medhdyd; II. dvena; pdrdhena; IV. candrdmdsd; vayodhdsd; 
purodhasd; V. -dakshasd; ddhasd. 

II. In some 1 8 instances, a form identical with the stem appears 
with the value of an instrumental. They may be taken outright 
as case-forms without ending (cf. pp. 476, 480, 535), or as text- 
corruptions arising from misconception of the irregular forms (cf. 
Bollensen, Orient und Occ. ii.481-2). The first six are pretty 
certain (cf. BR.) : vdcas, with divitmatd, i.26.2 ; with ndvyasd, 
ii.31.5 : vi.48.11 : viii.39.2 (cf. vi.62.5) ; vdcas, parallel with 
mdnmabhis, viii.63.1 ; yajnavacds, parallel with odanena, AV. 
xi.3.19; cdvas, parallel with krdtvd, RV. i.81.4. Ludwig trans- 
lates pdvas, x.23.5, by ' durch krafttat.' 

The rest are more or less doubtful : i.93.2, yd adyd vdm iddm 
vdeah saparydti, ' whoso to-day honors you here {iddm) with a 
song ;' i.32.8, mdno rtihdnd dti yanti d'pah, (' with a will ' 
i. e.) 'lustily rising, the waters overwhelm him;' x.109.1 ?, 
tdpa ugrd, Say., tapasd tdpanenogra udgUrno ''gnih; i.92.5, 
svdrum nd pepo viddtheshu anjdn, ' adorning the svdru as it were 
with beauty at the feasts.' 

Bollensen, I.e. p. 484, would explain mahd(s) or maho for mahd', i. e. mahdsd 
ini.22.11 (fdrmand): 1.121.11 (vdjrena): ii.34.12 (jyotishd): vii.25.1 (titi 1 ). In 
viii.3.6, fava i- might be taken similarly ; so Ludwig, Ueb. iLllS, ' mit kralt ;' but 
see G.s.n. In x.16.6, on account of divitmatd (m. or n.), Roth suggests that vded' 
(fern. 1) may be a corruption for vaco(—vaeasd)\ but cf. pp. 515-6. 

Dative Singtilae Masculine, Feminine, and Neuter. 

There are 74 masculines (from 38 stems) ; 4 feminines (from 4 
stems) ; and 225 neuters (from 27 stems). 

Masculines : abhishtipavase ; ukthdvdhase ; wuvydcase ; rtdpefase ; gdyatrd- 
vepase ; girvanase, 2 ; gwvdhase, 2 : ghwdcakshase ; jdtdveddse, 1 ; tddapase ; 
tddokase; tavdse, 11; dirghd-yapase ; -fravase ; duvdse; dyumndpravase ; nredk- 
shase, 2 ; niokase ; pitMnase ; pw&rdvase ; prthu-jrdyase ; -pd'jase ; prdcetase, 4 ; 
pratavase ; brdhma-vdhase, 3 ; mrkid- ; yajSd-, 2 ; yapdse ; rakshdse ; vipvdeak- 
shase ; v'ihdyase ; vedh&se, 1 1 ; pikvase ; satydrddhase ; sdhase ; sahdsracakshase ; 
svdtavase, 2 ; mdyapase. 

Feminines: tisha.se; dwvd'sase ; parjanyarelase ; vrdhdse. 

Neuters: apdse; apepdse ; dvase, 108; ojase, 5; cdkshase, 2; javdse; tdpase ; 
tdmase ; tdrase; ddkshase, 3; dhd'yase, 15; dhruvdse; ndmase; pd'jase, 2; 
prdyase, 2 ; psdrase, 2 ; mdnase, 4 : rdkshase, 3 ; rd'dhase, 25 ; retase ; vdcase, 3 ; 
vdrcase, 2 ; pdvase, 14 ; pr dvase, It ; sdhase, 1 ; svdyapase ; hdrase. 

Transition-forms (see pp. 550-557) : 1. jard'yai; II. pdrdhdya; 
krdnddya; tdrdya; card'yai; III. djarasd'ya; IV. vayodhdse; 
varoodhdse; dravinoddse; V. ddkshase; dhruvdse; vrdhdse. 

Ablative Singular Masculine, Feminine, and Neuter. 

I. There are 13 masculines (from 6 stems) ; 5 feminines (from 
one stem) ; and 107 neuters (from 24 stems). 



Ab.s.m.ln.] Noun- Inflection in the Veda. 563 

Masculines : a/rddh&sas ; jardsas, 2 ; tav&sas, 3 ; pdrinasas, 2 ; rakshdsas, 4 ; 
s&hasas. Feminine : ushdsas, 5. 

Neuters:- dnhasas, 44; dnasas ; dndhasas, 4; apasas; arnasas ; enasas,^; 
okasas; djasas, 3; kshodasas ; jr&yasas ; t&pasas, 3; tamasas, 10; dravinasas ; 
p&yasas ; p&'jasas ; prdthasas ; bhdnsasas ; manasas, 9 ; rajasas ; rd'dhasas, 2 ; 
vedasas ; sddasas, 9; sdrasas ; s&hasas, 2. In vii.56.19, s&hasas is N.p.m. 

Grassmann's ablative dr&vinodas-ds (accent impossible), ii.37.4, 
is based on a misprint ; read drdvinodasd-s, N.s.m. 

II. In vi.3.1, diva pd' si tydjasd mdrtam dnhah, if we take dnhas 
from dnh, the accent is anomalous. I think it is an Ab. of dnhas 
without case-ending, for dnhasas; cf. BR. vii.1685. A similar 
haplographia is seen in irddhiai, i. 134.2, for irddhadhiai. Since 
the accent of ush-ds as an accusative p.f. (iiiC?: viii.41.3) is at 
best a rare anomaly, I would strike out the stem Ush and regard 
ushds as an A. p.f. without ending, and put ushds (i.69.1,9: vii. 
10.1 : x.8.4) in the same category, as a G.s.f. without ending. In 
like manner is perhaps to be explained the form sddas-pdti, 1.21.5 ; 
of. sddasas pdtim, i. 18.6. 

Genitive Singular Masculine, Feminine, and Neuter. 

I. There are 58 masculines (from 30 stems) ; 45 feminines (from 
one stem) ; and 304 neuters (from 35 stems). 

Masculines : dngirasas ; dndgasas ; apasas, 2 ; dpdkacakshasas ; apnasas ; 
dbhivayasas ; arakshdsas ; upamagravasas ■ krshnajanhasas ; jdtavedasas, 6 ; 
tav&sas, 6 ; ddn&apnasas ; dirghagravasas ; dviidrhasas, 2 ; nrcdkshasas, 5 ; pari- 
dveshasas ; parinasas, 2 ; pwnibhdjasas ; prfhufrdvasas, 2 ; prdcetasas ; prdmdh- 
asas ; bhd'tvakshasas : raksh&sas, 5 ; vicetasas ; vidmand'pasas ; vifvdmanasas ; 
vr'kadvarasas (or A.p.) ; vedhdsas, 7; sabd'dhasas • svayofasas. 

Feminines: ushdsas, 44; ushd'sas, x.39.1 (seep. 546). 

Neuters: anhasas; dnasas ; dndhasas, 56; apasas; ay asas, 2; drnasas, 4; 
dvasas, 8; 6kasas,2; 6jasas,3; gdvarnasas ; tamasas, 10; tyqjasas,2; d'dkshasas ; 
dravinasas, 3; ndbhasas; n&masas, 3; pay asas, 2; pray asas; manasas, 12; 
manasas ; medasas, 3 ; rajasas, 41 ; rdpasas, 2 ; rd'dhasas, 19 ; rd'dhasas-rddh- 
asas; retasas, 4; vdcasas, 10; vdpsasas ; vdyasas ; vdrpasas, 4; vd'sasas ; 
( dvasas, 19; favasas, 16; prdvasas, 4; sddasas; sdrapasas; sdhasas, 13; sahasas, 
47 ; hdrasas. 

In i.102.6 we read pratimd 'nam Sjasd'thd, p. -asd dtha, and 
Benfey, Abh. xix.258, sees an instance of elision and crasis for 
djasas, as in verse 8. In i.62.9, silniir dddhdra cdvasd suddnsdh, 
p. -d, he sees loss of visarga and a metrical lengthening in the 
eighth place; Say., cavasah . . putrah. So in i.27.2, stintih 
cdvasd; S&y., balasya putrah. 

Contract form : nrmdnds, x.92.14. 

Transition-forms (see pp. 553-557) : II. cdrdhasya; niravdsya; 
IV. candrdm&sas; retodhdsas; V. ddJcshasas; doshdsas. 

II. In viii.3.6, indro mahnd" rddasi paprathae chdva i-, pdvas 
may pass for a G.s.n. without ending; 'Indra hath stretched out 
the heaven and the earth by the greatness of his might.' Say., 
cavasah . . . mahattvena. For sddaspdti, see Ab.s.n. Cf. the 
instrumentals, vdcas, etc. 



564 C. R. Lanman, [os-stems. 

Locative Singular Masculine, Feminine, and Neuter. 

There are 8 masculines (from 6 stems) ; 5 feminines (from one 
stem) ; and 45 neuters (from 18 stems). 

Masculines : r'j&nasi ; gavarnasi ; jdtdvedasi ; d&munasi ; prthufrdvasi • satya- 
fravasi, 3. 

Feminine : ushasi, 5. 

Neuters : dnJiasi ; diljasi, 2 ; apdsi, 2 ; avast, 2 ; d'gasi ; 6kasi, 2 ; krandasi ; 
cdkshasi ; tamasi, 1 1 ; namasi ; pdyasi ; manasi ; rdjasi, 6 ; rd'dhasi ; retasi, 2 ; 
fr&vasi; s&dasi, 8 ; sarasi. 

Contract form : updne ??, i.51.11. 

Transition-forms (see pp. 550-556) : I. dngire; II. hede; IV. 
eandrdmdsi; aprajasi. 

Vocative Singular Masculine, Feminine, and Neuter, 

There are 183 masculines (from 27 stems) ; 61 feminines (from 
2 stems) ; and 2 neuters (from 2 stems). 

Masculines: angiras, 11; ukthavdhas ; upamafravas ; kdrudMyas; girvanas, 
35 ; girvdhas, 4 ; jdtavedas, 66 ; devapravas ; dhrshanmanas, 2 ; nrcakshas, 2 ; 
nrmanas, 4 ; nodhas ; purilravas, 5 ; pracetas, 5 ; forahmavdhas, 3 ; mitramaJias, 
13; vdjapramahas ; vifva-eakshas ; -dhdyas ; -manas; vrshamanas, 2; vedhas, 
1 ; satyarddhas, 2 ; sudravinas, 2 ; sumahas, 3 ; svatavas ; suojas. For girvdhas, 
vi.24.6, Gr. reads girvavd'has ; see p. 353 med. 

Feminines: dhanas, 2; uslias, 59; tigmatejas, AV. vi.63.2. 

Neuters: dravinas (ague), iii.'7.10; fardhas (md'ruta), v.46.2 — cf. p. 339. 

Transition-forms (see pp. 550-556) : I. dpe; II. sdpratha; IV. 
candram&s; vayodhtte; dravinod&s, 6. 

The TS. at i.5.11 3 reads praceto rd'jan : the original passage in 
the Rik, i.24.14, reads pracetd rd'jan, p. pra-cetah (Prat, iv.13). 
In view of this irregular sandhi, we may take prdcetd as voc. in 
vi.5.5, sd mdrtieshu amrta prdcetd rdyd', although the pada has 
prd-cetdh. The accent may be due to a misconception of the 
form. Cf. Benfey, Abh. xix.260. 

Nom., Ace, and Voc. Dual Masculine and Feminine. 

I. There are 114 masculines in -d (from 54 stems); and 27 fem- 
inines in -d (from 11 stems). The vocatives are included with 
the nominatives. 

Masculines : apasd ; arepdsd ; dpiheshasd ; iyacakshasd ; ukthavdhasd ; wucak- 
shasd; krshtiojasd ; gambhfracetasd ; gopartnasd; jdtdvedasd (vii.2.7); tddokasd ; 
tavdsd ; navedasd ; nrcdkshasd ; nrvd'hasd, 2 ; pwud&nsasd, 4 ; purubhojasd ; 
pAtddakshasd, 4 ; prthupakshasd ; prdcetasd, 6 ; pramahasd, 2 ; bodhinmanasd ; 
matavacasd ; manojavasd ; ywjnavdhasd, 2 ; y apasd, 3 ; rifd'dasd, 4 ; vatsapracet- 
asd ; vicetasd, 2 ; vipakshasd ; vipravdhasd ; vifva-bhajasd ; -vedasd, 4 ; vedhasd ; 
vienasd; preshthavarcasd ; sajdshasd, 21; s&manasd, 6; samdndvarcasd ; samok- 
asd, 5; sdvayasd, 2; savedasd; sindhuvdhasd ; suddnsasd; sv/prayasd ; swddhasd; 
suretasd ; suvd'easd ; suavasd ; hitdprayasd ; hiranyapepasd, 2 ; further, camdrd- 
rnasd ; sftrydcandramasd ; tofd'sd (see p. 545). 

Feminines: anehdsd; uruvydcasd; bh&'ri-retasd, 3 (and AV. viii.9.12); -varpasd; 
vipvapepasd ; sacetasd ; sdmanasd ; suddnsasd, 2 ; swpepasd, 5 ; ushdsd, 4 ; ushd'sd, 
p. &', 2 : ndktoshd'sd, p. &', 5 ; see page 546. 



N.A.V.d.m.f.] Noun-Inflection in the Veda. 565 

The late verse, RV. x.14.11, has very properly the modern form 
nrcdkshasau at the end of b. For this, the Atharvan, perhaps 
with affected archaism, reads nrcdkshasd, xviii.2.12. 

II. There are 7 forms in -au : (masc.) arepdsau; nrcdkshasau; 
vipvdpardhasau; and, silrydcandramdsau, 2 ; — (fem.) yapdsau; 
ushd'sau, p. &', see p. 546. 

The Atharvan, at xiv.2.16, modernizes vienasd of the Rik (iii. 
33.13) to vienasau. 

Contract form : puruddnsdf, vii.73.1. 

Transition-forms: I. sajdshau; ushk, ; II. advesh'e; vidradM; 
IV. candrdm&sd; su'rydeandram&'sd; -au. 

Nominative and Accusative Dual Neuter. 

Here belong : dndhasi; krdndasl, 3 ; jdnasi; nd'dhasi; pd'jasi; 
rdjasi, 9; vdcast; from AV., ndbhasi, v.18.6 ; 20.7: xi.5.8,11 : 
xii.3.5,6: xvii.16; pdkshast, viii.8.22; from TS., drddhast, iii.2.2". 

Instrumental and Genitive Dual Feminine. 

Transition-form : I. ushd'bhydm, p. 551. 
Transition-forms : VI. rddas-os; aksh-ds ?; p. 558. 

Nominative and Vocative Plural Masculine and Feminine. 

There are 287 masculines (from 92 stems) ; and 94 feminines 
(from 17 stems). Vocatives are counted with nominatives. 

Masculines : dgiraukasas ; dngirasas, 19 ; acetdsas ; acoddsas ; adbJmtainasas , 
anavabhrdrddhasas, 4 ; dndgasas, 1 ; anehasas, 3 ; apdsas, 6 ; arepdsas, 7 ; dvicet- 
asas ; aceshasas ; dcvarddhasas, 2 ; dsdmicavasas ; dhandsas ; ukthdvdhasas 
•hdojasas; kshetrasd'dhasas ; gambhirdvepasas ; ghrshvirddhasas ; ghordvarpasas 
2 ; citrdrddhasas ; tddokasas ; tavdsas, 2 ; tuvishvandsas ; ddmunasas ; duvdsas 
dvib&rhasas; dhrshnticyasas ; ndvedasas, 2; nrcdkshasas, 6; p&tadakshasas, 2 
prthupd'jasas ; pracetasas, 14; pratijUtivarpasas ; prdtavasas ; prdtvakshasas, 2 
prdcravasas ; bdhuojasas, 2; bhaldndsas ; (bh&'rivarpasas, better A.p.l); mddhu- 
psarasas ; mrgayasas ; yajnavdhasas ; yacdsas, 5 ; yvtddveshasas ; rakshasas ; rip 
d'dasas, 14 ; rukmdvaJcshasas, 6 ; vdrunapeshasas ; vd'ta-ratihasas, 2 ; -svanasas , 
vicetasas, 4 ; vidmand'pasas ; vidyunmahasas ; vipravacasas ; mmahasas ; vicvd- 
dhdyasas; -mahasas ; -vedasas, 17; vishpardhasas ; vihdyasas; vrddhdcavasas, 2. 
vedhdsas, 1 8 ; ciJevasas, 3 ; creshthavarcasas ; sdcetasas, 4 ; sajdshasas, 28 ; satya- 
cavasas,2-, sabharasas ; sdmanasas, 8; sdmokasas ; savayasas ; sahdchandasas , 
sahdorapdjasas, 2 ; sucdkshasas ; sucetasas ■ suddnsasas, 2 ; supivdsas ; svpScasas, 
2; suprdcetasas ; sumdnasas, 5; sumahasas ; sumedhasas, 4; swd'dhasas, 2 
sti'racakshasas, 4 ; stdmavdhasas, 3 ; svdtavasas, 6 ; sudpasas, 5 ; sudpnasas 
svdyafosas, 3 • sudvasas, 2 ; hitaprayasas, 4. For sapsard'sas, see p. 546. 

Feminines: agnibhrdjasas ; anehasas, 18 ; apasas, 4; apsardsas; trshucydvasas 
dhdnvarnasas ; nrpecasas ; pracetasas ; prdsvddasas ; rnddhuarnasas ; yacdsas, 2 
vicetasas ; supecasas, 2 ; su'dadohasas ; sudpasas ; svdyacasas ; ushdsas, 41 
ushd'sas, p. #', 15, see p. 546. 

In i.79.1, I would read p&cibhrdjds\a\ ushdso ndvedds (cf. p. 
345 med.), and take all three words as N.p./em. (cf. p. 362). 
' Bright gleaming are the dawns, his companions.' 

Contract forms: (masc.) ndvedds, i.165.13; dngirds, i.83.4 ; 
dndgds,vii.81.7; 97.2: x.36.12 ; anehd'sf, x.61.12; sajdshds, 19 
times; (fem.) medhd's, Val. 4.9; djoshds, i.9.4 ; mrd'dhds, iii. 
33.12 ; ndvedds, i.79.1. 

vol. x. 75 



566 C. R. Lanman, [as-stems. 

Transition-forms (see p. 550-557) : I. dpd's; II. pepds; IV. 
vayodhdsas; suprajdsas; divdkshasas; V vegdsas; pdrivepasas; 
sabharasas; tuvishvandsas; vd'tasvanasas. 

Nominative and Accusative Plural Neuter. 

Here belong 284 forms (from 49 stems). 

Enumeration : Anhdnsi, 2 ; AnJednsi ; Andhdnsi, 6 ; apdnsi, 13 ; arqdnsi, 9 ; 
avdnsi, 4; d'gdhsi; endnsi, 2 ; 6kdnsi,2; ojdnsi; kdrdnsi; chanddnsi,i; jAvdhsi 
jraydnsi, 4 ; tamdnsi, 25; tvakshdnsi ; dahsdhsi, 3; duvdnsi, 2; dveshdnsi, 15 
pAydnsi, 6 ; pd'jdnsi, 2 ; pepdnsi ; praydhsi, 14 ; bhd'sdnsi ; mdndnsi, 6 
m&hdnsi, 3; rakshdhsi, 20, rdjdhsi, 39; rApdnsi, 4; rd'dhdnsi, 12; retdnsi, 2 
rddhdnsi, 2 ; rohdnsi; vakshdnsi; vAcdhsi, 12 ; vAydnsi, 4; vardnsi,5; vArivdnsi; 
varpdnsi, 3 ; vd'sdiisi ; pardhdnsi, 3 ; pdvdnsi, 7 ; pra»<2»m, 18 ; s&ddnsi, 3 ; 
sArdnsi, 6 ; sahdnsi, 4 ; skAndhdnsi ; heddmi ; hvArdnsi, 4 ; from AV., rakshdnsi, 
vocative, xi.10.1. 

Accusative Plural Masculine and Feminine. 

I. There are 51 masculines (from 21 stems) ; and 41 feminines 
(from 16 stems). 

Masculines : Ahgirasas ; Andgasas, 5 ; anehasas ; Apracetasas, 2 ; arddhasas, 3 ; 
urucAkshasas ; tuvird'dhasas, 2; dhvarasas; pAtAdakshasas ; prAcetasas, 2; &rft<$c- 
chravasas ; yajnavanasas ; yapasas ; rakshasas, 20 ; rep^'dosas ; vimahasas ; w{- 
shpardhasas, 2 ; (vr'kadvarasas, G.s.m. ;) vedhasas ; supipasas ; wrd'dhasas, 2 ; 
sttopwasos. 

Feminines : ajavasas ; anapn&sas ; ameh&sas ; apasas ; apsarAsas, 2 ; aridhdy- 
asas ; gharmasvarasas ; tadapasas, Khila ; bh&'rivarpasas ; yapAsaa ; vd'ja-dravin- 
asas ; -pravasas ; vipva-dohasas ; -dhdyasas ; supepasas ; ushasas, 25 ; for ushd'sas, 
x.39.1, see p. 546. 

Contract forms: (masc.) dndgds, vii.60.1 ; 66.4; sumedhd's ?, 
vii.91.3 ; (fern.) ushd's, ix.41.5. 

Transition-forms (see pp. 551-557) : I. dndgdn; II. pdrdhdn; 
IV. idaprajasas; kshdsasf; V. gharmdsvarasas. 

II. In iii.6.7: viii.41.3, I take ushds as A.p.f. without ending 
(p. 563). Perhaps also the best solution of the vexed passage, 
i. 11 2. 18, aiigiro m-, p. angirah, is to read dngiras, taking this as 
an A.p.m. without ending, for dngirasas. The strange form was 
probably mistaken by the copyist for a voc, and so left un- 
accented. See p. 342-3. 

Instrumental Plural Masculine, Feminine, and Neuter. 

Masculines : agnitdpobhis; dngirobhis, 10 ; svdyapobhis. 

Feminine : svdyapobhis (IMS'). 

Neuters (127, from 22 stems) : drnobhis; dvobhis, 17 ; ojobhis; 
tdpobhis ; tdmobhis ; tdrobhis, 2 ; ddnsobhis, 2 ; dv'eshobhis ; 
dhd'yobhis; ndbhobhis; ndmobhis, 45 ; pdyobhis, 5 ; prdyobhis, 
2 ; mdhobhis, 13 ; rdjobhis, 3 ; rd'dhobhis; vdcobhis, 13 ; vdyo- 
bhis; vdrobhis; pdvobhis, 2 ; prdvobhis, 6 ; sdhobhis, 7. 

Transition-form : medhd'bhis, p. 552. 

Forms after the analogy of the dental stems. Here belongs 
itshdd-bhis, i.6.3. For convenience of reference, the other similar 
forms are collected here : svdtavad-bhyas, VS. xxiv.16 : Katy. 



I.p.m.f.n.] Noun- Inflection in the Veda. 567 

C*. v.1.16 : Kath. xxxvi.l (regular, svdtavobhyas, Qat.Br. ii.5.1 14 ) ; 
mdd-bhis, -bhyds (p. 497) ; ad-bhls, -bhyds (p. 483) ; at-su /, cited 
by Prof. Ludwig, Rigveda, iii. p. ? . Cf. Pan. vii.4.48 ; Weber, 
Beitrdge, iii.387 ; Brugman, Zeitsch. xxiv.70. 

Dative and Ablative Plural Masculine and Neuter, 

Datives masc. : dUgirobhyas, 4 ; -bhyas, Gr. -bhias, i.139.7 : 
viii.52.3 : ix.62.9. 
Datives neut. : rdkshobhyas; vdyobhyas, AV. vi.10.2. 
Ablative neuter : dvishobhyas. 
Transition-forms : I. dpdbhyas; apsard'bhyas; pp. 550, 551. 

Genitive Plural Masculine, Feminine, and Neuter. 

Masculines : dngirasdm, 8 ; ddbhutainasdm ; apdsdm, 3 ; d- 
stheyasdm; tavdsdm ; dasmdvarcasdm ; mahd'manasdm ; yap- 
dsdm, 2 ; vedhdsdm; stdniavdhasdm. 

Feminines : apdsdm, 3 ; apsardsdrn; ndkshatrapavasdm; ush- 
dsdm, 22 ; for ushd'sdm, see p. 546. 

Neuters : ehdndasdm ; rddhasdm ; v'edasdm ; tdrasdm, AV. 
x.10.24. 

Contract form : vayd'm ?, see p. 552. 

Transition-forms (see pp. 551-558) : I. apsardndm; III. mahd- 
nasdndm; pracetasdndm; IV. purodhasdm; VI. sdraedm. 

Locative Plural Masculine and Neuter. 

I. Neuters: -d'dhassu, x.49.10 ; rdjassu, vii.34.16: viii.66.5 : 
x.43.8 ; vdkshassu, i.64.4 ; 166.10: v.54.11 : vii.56.13; prdvassu, 
iii.37.7 ; sddassu, vii.85.3 ; from AV., vdyahsu, iii.21.2. See 
Whitney to Ath. Pr. ii.40. The Rik-forms are all written with 
-hsu in Mtlller's first and second editions and in Aufrecht's first ; 
and with -ssu in Aufrecht's second; cf. p. 573 end. The ss makes 
the syllable long by position ; as in vii.85.3a. 

Transition-forms : I. apsard'su, p. 551 ; II. purupfydsu, p. 554. 

II. In AV. vi.35.2, agnir (na d'gamad) uktheshu dnhasu, 
Agni's presence is invoked not only at their ' festal songs,' but 
also 'in adversity.' At any rate, I can make nothing of dnha-su 
but a L.p. of dnhas- (cf. Ath. Pr. iv.32). If this is admissible, 
we may also take apd-su, viii.4.14, for apds-su, 'bei den emsigen.' 
Say., asmadiyeshu karmasu . . . sakdro lupyate. Bradhndm 
refers to the soma ; and if we took apd-su for ap-sti (cf. p. 484), 
the result would be an uninvitingly thin drink for Indra. Cf. 
md-su- for mdss'h, p. 497. See p. 554, II. end. 

The same peculiarity is seen in several Zend forms : usha-hva, 
Vend, xxi.3 ; areza-hva, ibidem, bis; hazanrd-tema-hva-ca, Yt. 
xv.53 ; dza-hu, Yac. x. 17; dza-hil, Yt. xxiv.51. The last two 
forms agree remarkably with dnha-su, above. Examples of the 
regular form are : temd-hva, Yt. xxii.33 ; uzird-hva, Vend, xxi.3. 



568 C. B. Lanman, [is and ws-stems. 

STEMS IN IS AND US. 

The stems in is and us show almost no feminine forms. The 
stems in is include about twelve neuters (10 oxytones) ; the mas- 
culines, excepting d'rnis, are exclusively compounds of these 
neuters. The forms are enumerated under category A. 

The stems in us include neuters and masculine compounds 
thereof; and also a number of primary masculines: as, n&hus. 
The forms are enumerated under category B. 

Stems in is and us distinguish strong and weak cases only in 
the N.A.p.n., where the thematic i or u is lengthened and nasal- 
ized, precisely as with the as-stems. 

The relations of stems in is and us to equivalent vocalic themes 
demand a treatment, which, as far as it goes, is parallel with that 
of the preceding chapter (p. 548). I shall therefore discuss : 

L Transitions of the stems in is and us to the declensions in i 
and w (cf. II. p. 553). 

II. Transitions to the a-declension (cf. III. p. 554). 

III. Transitions of stems in i and u to the s-declension (cf. V. 
p. 556). 

I. Transitions to the Declensions in J and U, 

Just as a neuter like h'edas (p. 553) passes into the a-declension 
(hbda-m), so here a neuter like arcis passes into the i-declension 
(arcdyas). 

A. 1. The stem pocis and its compounds occur frequently ; see 
BR. The form pods is sometimes marked as a neuter by an 
adjective, if nom. (e. g. viii.23.4) ; and by the construction, if ace. 
In vii.S.5d, however, didd'ya pocir d'hutasya vr'shnah, it may 
easily be felt as a N.s.masc. ; so also in vi.64.25 and vii.16.3a/ 
in x.16.46, tdrh te $ocirS tapatu, tdrh te are'v-h; and in AV. i.25.2. 

As transition-forms may be regarded : (yd's te) pocdyas, AV. 
xviii.2.9 ; pdvaka-poce, 2 ; bhadra-poce, 4 ; pukra-poce; see p. 
390 top. In vi.64.25, ut te pocir bhdndvo dyd'm apaptan, Gr. 
proposed the transition-form $oc'es ( Ueb. i.581) ; but this is un- 
necessary. Translate : ' Up sped thy light — thy beams unto 
heaven.' 

2. In x.l6(842).4 : AV. i.25.2 : xii.1.51, arcis may be N.s. 
neuter of arcis, or N.s.m. of arci (arcb-s). Such a form is the 
point of departure for the two transition-forms : arcdyas, RV., 
14 ; arci-bhis, vi.48.7 : v.79.8. In the latter passage it is called 
out by the formal parallelism of the verse : sdkdrh siX'ryasya 
ra$mibhih \ pukraih $6cadbhir arcibhih. In ii.3.2, suarcis {ndrd- 
$dnsas) may just as well be referred to suarcis as to suarci (BR., 
Gr.). 

3. The stem krav-is is a correct formation ; with it cf. d-kravi- 
hasta, v.62.6. 

4. With jy6tis cf. jyoti, Kern, Translation of Brh. S., p. 135, 
note 2. 



Transitions. L] Noun-Infleclion in the Veda. 569 

B. 5. The stem cdJeshus is well authenticated ; but its nom. 
cdkshus might be felt as a N.s.masc, cdkshu-s, in x.8.5 and 16.3. 
Such ambiguities account for the transition-forms : cdkshos (see 
p. 410 top) and sahasra-cahsho (p. 412 end). 

6. In like manner the stem tdpus is amply supported. The 
ambiguous N.s. tdpus, vii.104.2, is referred by Gr. to stem tdpus; 
but by BR. to tdpu. So tdpus (agnis), ii.4.6, may just as well be 
referred to tdpus, although both lexica agree in assigning it to 
tdpu. There remains only one form, tdpos, ix.83.2, to be ascribed 
to false analogy. 

7. The form dhdnus, ' a bow,' is generally neuter — stem dhdnus. 
In the late hymn vi.75, verse 2, it may be masc. (dhdnu-s) ; while 
in Hit. Procem. 22, dhanu-s nirgunas, it must be masc, i. e. a 
form of transition to the M-declension. 

8. The Rik has 11 forms from vanfis, and but two from vanu: 
viz., -um and -u'n. The latter are probably transition-forms ; but 
a nom. vanu-s does not occur. 

9. At iii.699, BR. refer the an. A. duhpd'sus, N.s.m., x.33.1, to 
duh-pd'su; but at vii.170, correctly, to duh-pd'sus. Cf. Gr. 

io. The N.s.m. vidus, i.71.10: vii.18.2, is referred to vidus. It 
might be referred to vidu; cf. rip-u, vidh-u. 

11. Probably tdru has nothing to do with tdrus. The an. A. 
tdrubhis in the late and corrupt hymn v.44, verse 5, seems to me 
to be a vague reminiscence of tdrobhis, ii.39.3, q.v. 

12. The stems dyu, 'active, lively,' and d'yus, 'life,' are of 
course independent formations. The latter is of frequent occur- 
rence in the Rik (92 forms), and is supported by the compound 
stems kshitd'yus, citrd'yus, dirghd'yus, purvdyus ?, patd'yus (all 
posited by Gr.), sarvd'yus, gatdyus, and, as I think, by vipvd'yus. 

I do not believe that there was another stem d'yu, ' life,' of 
independent formation. BR. refer d'yus, i.66.1, to d'yus; Gr., to 
d'yu. In x.17.4, also, Grassmann's d'yu-s can just as well be taken 
as a neuter. The Rik has but one form not referable to d'yus, 
viz., d'yuni (3) ; and I therefore regard this as a transition-form, 
and d'yu as a secondary stem. 

But in support of d'yu, ' life,' BR. (i.678) quote the compounds 
adabdhdyu, hshitdlyu, dirghd'yu, vipvd'yu, vrddhd'yu, and sarv- 
d'yu. Even these, I suspect, fail to establish a stem d'yu. 

a. The air. A. kshitd'yus, N.s.m., is referred by G-r., and even by BR. s.v., to an 
«s-stem. 

6. The N.s.m. sarvd'yus, "VS. xxxviii.20 : Qat.Br. xiv.3.1 18 : Ait. Br. ii.T, is 
referred by BB. to " sarvdyu or -us;" better, I think, to -us. 

c. Dirghd'yush-am occurs iv.15.10; in verse 9, dirghd'yus is a N.s.m. of 
ambiguous stem; so in x.85.39 and VS. xii.l00a; and in TS. xii.lOOc, N.s.f. 
Such forms pave the way to the true transition-yoeatire, dirghdyo, viii.59.7. 

d. Entirely similar is the V.s.m. (dgne) adabdhdyo, VS. ii.20. 

e. In i.10.12, vrddhd'yu-m is merely a air. A. 

/. The N.s. vifvd'yus occurs in the Bik 18 times, and in VS. xxxviii.20 parallel 
with sarvd'yus. Since the support for a real stem d'yu is so weak, I am tempted 
to refer this N.s. to stem vifvd'yus, and to regard -u (5), -um (3), -ave, and -os as 
forms of transition. 



570 0. R. Lanman, [is and ws-stems. 

In i.31.5d, ekdyur agre vipa dvivdsasi (BR. -ti), the ott. A. ekdyur may be taken 
as voc. or as nom. of ekdyus (eka + d'yus), 'having unique vigor,' i. e. ' of incom- 
parable might.' BR. (i.1087, q.v.) refer it to ekdyu, i. e. eka + dyu. In no case is 
it to be taken, with Gr., as a compound of d'yu, 

II. Transitions to the jI-Deolension. 

The transition results in forms from a longer vocalic theme ; 
cf. pp. 554-5. 

B. 1. The stem ndhus is a primary formation. From some of 
its ambiguous forms has arisen by way of transition the secondary 
stem ndhusha. Thus in x.99.7c, sd nrtamo ndhusho , smdt sujdtah, 
nd- may be G.s.m. of the collective ndhus : ' He, the manliest of 
the neighborhood,' i. e. ' of our allies ;' or N.s.m., ndhusha-s : 
'He, the manliest ally.' The same ambiguity prevails in i.122.10 
(Gr., A.p.m. ; Ludwig, N.s.m.). Corresponding to ndhusha-s is 
made the G.s. ndhushasya, i.31.11 : v.12.6 ; and the L.s. ndhushe, 
viii.46.27. The stem ndhusha survives the Vedic period. 

2. The stem mdnusha never had any real existence. In i.26.4, 
mdnush-as is N.p.m. ; but it may easily be felt as a N.s.m., 
mdnusha-s. Mdnushas in x.11.5 is taken as G.s.m. by Ludwig 
and Gr. Ueb. ; and as N.s.m. by BR. and Gr. Wb. In x. 
49(875). 7, it is by no means an a-form. In the Rik there are but 
two real transition-forms: mdnushdya, i. 117.21 : x.65.4 ; -asya, 
i.31.11 (cf. ndhushasya). 

3. The an. X. vdpushdya (=vdpush-e), iii.2.15, is perhaps a false 
form called forth by the metre. 

III. Transitions to the /S-Declension. 

A. 1. BR. derive the an. A. surabhish-tama, i.186.7, from 
"surabhis, nom. of surabhV Cf. indras-vantam, iv.37.5. 

2. The compounds with turn- are numerous. With tuvis- we 
find only tuvish-mant (13) and tuv'tsh-tama (3). One occurrence 
of the latter, curiously enough, is in the verse next before sura- 
bhish-tama. The AV. pada reads tuvi-tama, Prat, iii.96. Since 
the suffix is is regularly accompanied by guna (cf. krav-is, hay-is, 
roc-is, vart-is), I am the more inclined here to regard tuv-i (cf. 
y&'-yuv-i) as the original formation. Otherwise, we should expect 

B. 3. Since a suffix su is at best very doubtful, I regard 
dhdhshu as an adj. from a desiderative without reduplication, 
dhdksh-u, for di-dhaksh-&, like tr'ts-u for ti-trts*u. This occurs 
in x.115.4, dhdkshos; and in ii.4.4, dhdkshos — sanhita ddkshos 
(Prat, iv.41). And since a suffix sus, or us added to a desidera- 
tive stem (odkshus=ca-kas-us), is unheard-of, I would consider 
the form dhakshush-as, i.141.7, sanhita da- (Pr. iv.41), as due to 
false analogy. 

4. Perhaps the stems mdnu and manus are independent forma- 
tion's ; cf. rndnavd and md'nusha, manuvdt and manushvdt. It 
is to be noted, however, that while mdnu has a tolerably complete 
inflection (-«s, 15 ; -urn, 3 ; -und; -ave, 36 ; -os, 8 ; -au, -au, 5 ; -avas, 
4; -undm; sum, 73), only 3 forms of mdnus are used (-ushd, 



Transitions. III.] Noun- Inflection in the Veda. 571 

2 ; -ushe, 8 ; -ushas, 47) ; and that while mdnu remains in use in 
the later language, mdnus becomes obsolete. Perhaps, then, the 
Vedic forms of mdnus are due to false analogy, mdnus being felt 
as N.s.m. of a stem mdnus. Cf. also vipvd-manush-dm. 
5. With dprdyu (root yu), cf. dprdyush-e, an. A., i.127.5. 

THE FEMININE FORMS. 

A. Sole example : svd-pocis (rodasf), N.s., vi.66.6. 

B. Here belong : N.s., dghoracakshus (dpatighnt), x.85.44 ; 
cdhskus {devil), i.92.9 ; citrddyus (kanid), vi.49.7 ; hradkcakshus 
(pr^nis), x.95.6 ; A.d., tdpushd (cakriyd), ii.34.9 — see p. 391, 
C.2 ; A.p., gdvapushas, x.68.9. 

THE MASCULINES AND NEUTERS. 

Nominative Singular Masculine. 

A. Here belong : dhavis; hrshndvyathis ; citrd-pocis; durdka-; 
pukrd-/ stirndbarhis; svdrocis; suarcis (p. 568) ; from AY.,jivd- 
barhis, xi.7.7. 

B. Here belong : kshitd'yus; cdhshus (x.8.5) ; dirghd'yus, 2 ; 
duhpd'sus; vdpus, 2; vidds?, 2; vipvdtapcakshus; from AV., 
hshitd'yus, iii.11.2; sahdsrdyus, xvii.27; ddabdhacakshus, xiii. 
2.44 ; viparus, vii.56.4 ; sdrvaparus, xi.3.32-49Ws. 

Transition-forms (see pp. 568-571): I. poci-s; arc<b-s; cdhshu-s; 
tdpu-s ; dhdnu-s ; vidu-s ? ; dirghd'yu-s ? ; II. ndhusha-s ? ; 
mdnusha-s?; III. mdnus?. 

If, with BR. and Gr., we refer the form janu's, vii.58.2, to 
janus, then the long u is entirely anomalous. Does it not belong 
in the same category as prajanil', dhanu 1 (p. 402, 2) ? if so, we 
should add it on p. 406, N.s.f., C.2, janil'-s. 

Nominative and Accusative Singular Neuter. 

A. Here belong 320 forms (from 10 stems) : areis, 6 ; chadis; 
chardis, 19;jy6tis, 96; barhis, 86; vartis, 25; vydthis,5; pocis, 15; 
sarpis; havis, 65 ; hav'w-havis; from AV., Jcravis (viii.6.23) ; etc. 
^ B. Here belong 143 forms (from 8 stems) : d'yus, 76 ; cdkshus, 33 ; 
tdpus, 2; dhdnus, 6 ; pdrus ; pdrus-parus, 2 ; ydjus, 5 ; vdpus, 13 ; 
pd'sus, 5; from AV., drus (v.5.4) ; indradhanus (xv.1.6) ; etc. 

Transition-forms (p. 569) : I. vipvd'yu?. In i.89.9 and iii.49.2, 
Gr. would read d'yu, text d'yur (cf. p. 406); but see Muller, 
Translation, p. cxxxiii. 

Accusative Sin»ular Masculine. 

A. Here belong : -pocisham, with dgrbhita-, 2 ; ajird-; urdhvd-; 
citrd-, 2 ; dirghd/yu-; pdvakd-, 6 ; ptr'd-, 2 ; pukrd-, 2 ; pr'eshtha-; 
— citrd-barhisham, 2; vrktd- ; subarhisham; from AV., ' ddk- 
shindrjyotisham, ix.5.22, 24-26, 28, 31-36 ; hiranya-, x.9.6. 

B. Here belong : janusham, 3 ; dirghd'yusham ; purvd- 
yushamf; patd'yusham. 

Transition-forms (p. 569): I. vanum; vrddhd'yum, vipvdlyumf. 



572 G. B. Lawman, \is and ws-stems. 

Instrumental Singular Masculine and Neuter. 

A. Neuters (141, from 7 stems) : areishd, 14 ; kravishd; jy6t- 
ishd, 34 ; rocishd; pocishd, 34 ; sarpishd; havishd, 56. 

B. Masculines : tdpushd; ndhushd; vanushd. 

Neuters : d'yushd, 4 ; cdkshushd, 2 ; janushd, 20 ; tdpushd; 
pdrushd; ydjushd; patd'yushd, 2. 

Transition-form (p. 571) : III. mdnush-d ?, 2. 

Dative Singular Masculine and Neuter. 

A. Masculines : tigmd-pocishe ; pdvakd-; pukrd-, 3 ; vrktd- 
barhishe, 3 ; sttrnd-; rdtdhavishe; suhdvishe. 

Neuters : pocishe; havishe; jydtishe, AV. vi.61.1. 

B. Masculines : edkshushe-eakshushe; vanushe. 

Neuters : cdkshushe, 2 ; jamdshe ; vdpushe, 8 ; from AV., 
tdpushe, i.13.3 ; d'yushe, 20 times, as i.35.1. 

Transition-forms (see pp. 569-571) : I. vipvd'yavef; II. mdnu- 
shdya; vdpushdya; III. mdnush-ef, 8 ; dprdyushe. 

Ablative Singular Masculine and Neuter. 

A. Neuters: barhishas; havishas; jydtishas, AV. iv.1.5 ; 10.1. 

B. Masculines : ndhushas, 2 ; vdpushas. 
Neuters : janushas, 4 ; vdpushas. 

Transition-forms (see pp. 569-571): I. edkshos? ; III. rn.dnusb.-asf. 

Genitive Singular Masculine and Neuter. 

A. Masculines : pdvakd-pocishas; vrddhd-; vdsurocishas; vrk- 
tdbarhishas, 5. 

Neuters : kravishas, 2 ; chardishas; jydtishas, 4 ; barhishas, 
6 ; pocishas; pocishas pate; sarpishas, 2 ; havishas, 1 1. 

B. Masculines : cdkshushas; ndhushas, 2 ; vanushas, 3. 
Neuters: d'yushas; cdkshushas; tdpushas; tdrushas, 2; pdr- 

ushas; vdpushas. 

Transition-forms (see pp. 568-571): L poets?; tdpos; vipvd'- 
yos ?; II. ndhushasya; mdnushasya; III. dhakshush-as; mdnush- 
as f, 37. 

Genitive without ending (? cf. pp. 562-3) : d'yur nd prdn.6, 
i.66.1. 

Locative Singular Masculine and Neuter. 

A. Masculine : d'mishi, 2. Neuters : arcishi; jydtishi; tri- 
barhishi; barhishi, 44 ; havishi, 3. 

B. Neuters : d'yushi, 2 ; tdrushi; pdrushi; vdpushi. 
Transition-forms (see pp. 569-570) : I. d'yuni, 3 ; II. ndhushe. 

Vocative Singular Masculine and Neuter. 

A. Neuter : (deva) barhis, ii.3.4 : x.70.4. 

B. Masculine : 'ekdyus ?, i.31.5 ; see p. 570 top. 
Transition-forms (see pp. 568-569) : I. pdvakapoce, 2 ; bhadra- 

poce, 4 ; pukrapoce; sahasracaksho, AV. iv.20.5 : xix.35.3 ; dir- 



V.s.m.n.] Noun- Inflection in the Veda. 573 

ghdyo, RV. viii.59.7; adabdhdyo, VS. ii.20. These vocatives are 
especially interesting and instructive as illustrating the manner 
of avoiding unfamiliar forms; cf. pages 390 top and 412 end. 
One need only study the dictionary articles pdvakdpocis etc., in 
order to be convinced that they are true transition-forms. 

Nominative and Accusative Dual Masculine and Neuter. 

B. Masculine : jaydshd, 3. Here also belongs cdhshushd, ii. 
39.55 : ' Ye (who are) discerning as two eyes, come hither.' I 
take akshi' as fern., p. 391 med. 

Neuters : jantisM; cdkshushi, AV. ix.5.21. 

Nominative and Vocative Plural Masculine. 

A. Here belong: dgrbhita-pocishas ; ajird-; vrktd-barhishas, 
14; vrkta-, 3; su-; sddma-; sujyStishas; sujyotishas (x.89.15) ; 
svdrocishas. 

B. Here belong : ndhushas; vanushas, 6 ; vdpushas. 
Transition-forms (see pp. 568-571) : I. focdyas; aredyas; III. 

mdnush-as ?, 5 ; rnanush-as. 

Nominative and Accusative Plubal Neuter. 

A. Here belong: arci'nshi ; jydtinshi, 8; poci'nshi, 2; hav- 
i'nshi, 15 ; cf. AV. xi.5.13 : x.7.40 : v.27.1 : iv.26.4. 

B. Here belong : dlyunshi, 9 ; cdhshAnshi ; jawd'hshi, 3 ; tdp- 
tinshi, 2 ; vdpunshi, 8 ; from AV, ydj4nshi (e. g. v.26.1) ; pdr- 
tinshi (e. g. ix.6.1 : VS. xviii.3) ; etc. 

Accusative Plubal Masculine. 

A. Here belongs sujyStishas. 

B. Here belong : ndhushas, 3 ; van&shas. 
Transition-forms: I. vanH'n, p. 569; mdnush-as?, 2, p. 571. 

Instrumental Plural Neuter. 

A. Sole example: havirbhis, 12, and AV. xviii.3.63 ; 4.54. 

B. Here belong: vdpurbhis; dhdnurbhis, AV. v. 18. 8. 
Transition-form : I. arci-bhis, p. 568. 

Genitive Plural Masculine and Neuter. 

A. Neuters : jydtishdm, 2 ; hav'ishdm, AV. vi.15.3. 

B. Masculines : vanHshdm, 5. Neuters : janushdm, 2 ; vdp- 
ushdm, 2 ; from AV, cdkshushdm, v.24.9; ya.jushdm, xv.6.3. 

Transition-form: vipvd-manush-dm?, p. 57i. 

Locative Plural Neuter. 

A. Sole example : havishshu, ix.7.2 ; so Mtiller's first and 
second editions and Aufrecht's second. Aufrecht's first edition 
writes hav'thshu. Cf. p. 567. 

vol. x. 76 



574 C. R. Lanrnan, [Add. note 1. 



ADDITIONAL NOTES. 



1 . Dual Forms in -& and -au from Stems not ending in a. 

§1. The N.A.V.d.m. of a-stems ends generally in -d (1129 forms, 
from 358 stems) ; sometimes in -au (171 forms, from 87 stems). 
The first two sums require several corrections. The 15 forms in 
-d, p. -au (p. 341), should be added, of which 7 are from uncounted 
stems ; and purdhitti, r-, p. -d (p. 342 end), should be added, since 
the other cases like it are included. The correct sums are then : 
1145 forms, from 366 stems. 

§2. The circumstances of occurrence are given above, pages 
340-342, which see; cf. also p. 427. 

§3. Either of these two endings may also appear in the N. A.V.d. 
masc. and fem. of all non-a stems, excepting those in short \ and & 
of series A (pp. 366, 401), and those in long i of series B. There 
occur in the Rik 1246 dual forms in -d (from 354 non-a-stems) ; 
and 122 in -au (from 58 non-a-stems). 

On page 390 are given from stems in i, m., 1 3 forms in -A and in -aw; p. 391, 
i, 1, 14 and 0; p. 400, sakhdy, 6 and 1 ; p. 413, u, m., 21 and 0; f., 15 and 0; p. 
427. r, 176 and 10; pp. 431-3, go, dyo, vrshiidyo, 30 and 4; p. 449, rad. a, A, 27 
and 7 ; p. 458, ac, 8 and ; p. 460, c, 7 and 3 ; p. 461, eft, 1 and ; p. 464, /, 45 
and 3; p. 468, t, 14 and 2 ; p. 473-4, d, 16 and 1 ; p. 477, dh, 20 and 2 ; p. 480, 
n, 8 and 0; p. 482, p, 2 and 0; p. 485, m, 9 and 0; p. 488, r, 13 and 4; p. 491, 
c, 7 and 1 ; p. 496, *, 7 and 0; p. 500, ft, 11 and 2; p. 503, voc. root + 1, 14 and 3; 
p. 509, nt, t, 86 and 16; p. 513, vans. 13 and 1 ; p. 520, vant, mant, 25 and 2; p. 
528, 1 1, sasthd'odnd ; p. 537, van, man, an, 121 and 22; p. 544, in, 369 and 32; 
p. 564-5, as, 141 and 7 ; p. 571, 1. 11, tdpusM; p. 573, us, 4 and 0. The sums are 
1245 forms in -d and 123 in -aw. But sakshitau, i.140.3, ought to be counted as 
an d-form, p. 503. In i. 120.3, vidvd'iisd, occurring twice, was counted but once; 
while cakriyd, x.89.4, was counted already under the a-stems. The correct sums, 
then, are 1246 and 122. 

§4. The circumstances of occurrence coincide entirely with 
those of the duals from a-stems. Cf. also the locatives of t'-stems, 
p. 385. 

A. The ending -d is used 

1. at the end of a pdda (433 times) ; 

2. before consonants (670 times) ; 

3. before an initial vowel with which it is fused (99 times). 

4. It is never used before non-w-vowels with hiatus. 

B. The ending -au is used before vowels without hiatus in the 
form -dv (90 times). 

A. 1. Of the 433 forms, 130 are at the end of pdda b and 85 at 
the end of d or of the verse ; 148 are at the end of a and 70 at 
the end of e. 

a. When an d-form at the end of a or c is followed by a-, i-, or u- at the begin- 
ning of o or d, the concurrent vowels are regularly written as coalescing in the 
saihhitd; but are of course invariably to be pronounced with hiatus. For 



Add. note 1.] Noun- Inflection in the Veda. 575 

example, in v.73.7: vi.62.7 : vii.73.1, the following pdda begins with a-; in 
i.111.21; 159.1: v.68.4, with %-; in i.46.15; 117.21: ii.31.5: iu.35.3: vii.42.5; 
99.6, with u-. 

j3. A merely graphic peculiarity is the short -a at the end of an odd pdda where 
the next begins with r-. See Kik Pr. ii. 1 1 : Ath. Pr. iii.46. This is not a real 
shortening ; cf. Benfey, Gram. p. 294, note 6. The instances are : vicvdcambhuvd 
r-, p. -a, i.160.1 ; rtdvr'dhU, v.65.2 ; prdcetasd, x.36.2 ; for others, see p. 342 ad fin., 
p. 355 ad fin. 

y. Irregular is the sandhi of acvind, viii.9.9c, with evd, 9d: acvinevd; see Prat. 
ii.38. 

S. In i.l40.3a6, we have krshnaprutau vevije asya sakshitd ubhd' ta/rete abhi 
mdtdrd cicum, p. -au u-. Since two pddas never coalesce metrically, we might 
suppose, if this were an old hymn, that the oral tradition had sakshitd ubhd'. 
This would be written regularly sakshitobJid', p. -d a-. This instance is entirely 
unparalleled in the Rigveda, and is interesting for the criticism of i.140. 

e. There are but six instances in the Rik in which a dual -dv ends an odd pdda 
and the next begins with a vowel. They are: i.23.5, rtdvr'dhdv r- ; v.34.8, 
-cardhasdv a-; 75.3, bibhratdv a-; x.132.1, afvindv a-; from a-stems, x.184.2, 
devd'v a' ; i.2.8, -varundv r-. The first and last, if conformed to the canons of 
the old parts of the Rik. would read -a r- ; the rest -d-. Cf. p. 544. These facts 
are of importance for the history of the text. Hymn x.132 is unclear and metri- 
cally corrupt. In i.2.8, the unaccented rtdvrdhau is a text-error. 

2. In 670 instances -d stands before consonants in the interior 
of a pdda. The exceptional instances of -au before a consonant 
are given under B.2. 

3. In 99 instances -d coalesces with a following vowel in the 
interior of & pdda. In 51 instances the result is d: as, cdkshushd', 
ii.39.5; in 38, it is e: as, apvineddm, v.76.4; in 10, it is o: namely, 
i.116.3,19: iv.2.4: v.76.2: vii.44.1: viii.5.2; 22.12: x.125.1; 128.7; 
131.5. 

4. Limitations to rule A.4. 

a. The cases in which -d is followed by u- or &- in the interior of a pdda form 
a class by themselves. In the 10 passages just cited, the concurrent vowels are 
run into one syllable, as the metre demands. If the metre requires them to be 
kept apart as two syllables, then the Riksanhita writes them with hiatus (-a «-), 
and the pada reads always -au «-. See Rik Pr. ii.9. The cases number 40 and 
are: i.108.3; 116.5; 161.10; 180.1; 182.7: iii.4.6: v.43.9; 63.3: vii.60.9; 72.4- 
73.2: viii.35.13a.14a; 35.1-21: x.14.12; 65.10; 66.76; 90.11: viii.9.UMs, 
ydtarh chardishpd' utd nah paraspd' bh&tdm jagatpd" utd nas tan&pd'. For sakshitd, 
see l.<5. The omission of afvinau hi the porfa-extraets to vii.72.4 and 73.2 of 
Aufrecht 2 is due to a mere inadvertence. 

/3. Of the 1246 (J-forms there remain four. These occur before a vowel with 
hiatus, but in such parts of the text as to confirm our rule rather than break it 
down. The first is vidntd pd'ntd aur r ij6 huvddhyai, sanhita pd'ntaut-, p. -d au-, 
i.122.4. This hymn is referred by Kuhn, Beitrage, iv.212, to a later redaction. 
The second is in the late and wretched bombast, x.61., verse 4, divo ndpdtd acvind 
huve vdm. The third is in x.22.6, adha gmantd ufand prchute vara. The hymn 
(noticed before at p. 341 end) has an antique look, but abounds in metrical corrup- 
tions: cf. the false accent vedid'ndm, for ved-, 14. In vii.69.3, if we read sudcvd 
yacdsd d' ydtam arvd'g, the hiatus is forbidden by our canon, and the caesura is 
intolerable. A good solution is : sudcvd [no] yacdsd' ydtam arvd'g. 

The hiatus is only apparent in viii.18.16c; the pdda is a syncopated jagatt: 
thus, dyd'vdJcshdmdre o asmdd rdpas krtam, — ' — ' — o — ' <~ —' ~ — . 

B. The ending -au is found 122 times. It occurs as -dv in the 
interior of & pdda 86 times, i. e. 70 per cent of the whole number; 
cf. p. 342. This is its normal use. 



576 G. R. Lanman, [Add. note 1. 

1. Four times it occurs at the end of an odd pdda as -dv before 
a vowel; see A. I.e. 

Four times it occurs as -au at the end of an even pdda or of 
the verse: x. 14. 115 (late) : 1.120.2c (corrupt) : ix.70.4.o\- x.65.5a". 
These are exceptions to A.l. 

2. The ending -au occurs 23 times in the interior of pdda before 
a consonant, and 5 times at the end of an odd pdda before a 
consonant. 

If any one will take the trouble to examine these passages, he will find in most 
of them other signs showing that they belong to decidedly late parts of the Rik- 
text: thus, 13 (about one-half) are in the tenth mandala ; 10 (more than one-third) 
are found in Grassmann's Anhang; and 15 (the first 12 and the last 3 as arranged 
below) coincide with those discussed on p. 341 top, which see. 

The passages are as follows: iii.33.13c (verse 13 is a patched-on fragment); 
53.17a (verses 8-24 are a separate late fragment as shown by contents): v.36.6a: 
vi.60.14c (13-15 are isolated fragments of different metre from the rest): vii.50.2e- 
(of real Atharvan character) : ix.112.4c (a late obscene addition): x.l4.11a (hounds 
of Yama — see p. 427); 27.20a (does not belong to the hymn proper): 85.116, 18i, 
42c (Surya's wedding) ; 117. 9d (backsheesh soug) ; i.95.7o; 140.3a (later 

redactiou?): ii.43.lc (late, as shown by sarv&tas, p&nya, etc.): iii.54.16c; v.37.56: 
vii.70.4c; 93.2c: x.lOl.lOd, 11a; 184.2c (a real Atharvan hymn— cf. p. 544); 
190.3a (a late cosmogonical hymn): — at the end of an odd pdda, iii.55.11c 
(dark and mystical) : x.75.4c (to the Rivers; Ganges mentioned) ; ii.40.lc: 

x.l4.10a, 11a (late). 

C. Duals from a-stems are sometimes found with short -& in 
both texts or in the sanhita only ; see p. 342. This never happens 
with non-a-stems. For prdcetas-d etc., see A.l. /J. For gmdntd, 
p. -td, see p. 509. 

§5. The Atharvan has in passages peculiar to itself 

A. 55 forms in -d from non-a-stems. The stereotyped apvind 
counts for 35 of these forms. The AV. has -d in two Rik-verses, 
where the Rik has -au: v.25.3, apv'md (see p. 544); xviii.2.12, 
nrcdkshasd (see p. 565). 

B. 105 forms in -au from non-a-stems. The AV. modernizes -d 
of the Rik to -au in the following Rik-verses : xviii.1.54, mddantau 
rd'jdnau (pp. 509, 537): xiv.2.16, vienasau (p. 565). For asutr'pdv 
«-, see p. 482. 

2. The relative Frequency of ancient and modem equivalent 
grammatical Forms as a Criterion of the Age of different 

Vedic Texts. 

§1. We have here to deal only with the four most important of 
the many pairs of grammatically equivalent terminations. In 
each pair, the archaic ending is put first ; the modern ending, 
second. They are : 

I. -d and -au; see N.AV.d.m.f., pp. 340, 574. 
II. -dsas and -ds; see N.V.p.m., p. 344. 

III. -a and -dni; see N.AV.p.n., p. 346. 

IV. -ebhis and -ais; see I.p.m.n., p. 349. 

Only the first pair is used with non-a-stems. The feminine 
forms in -dsas, p. 362, the 27 forms in -d and those in -dni from 
stems in an, p. 539, are very properly left out of account ; so too 



Add. note 2.] Noun- Inflection in the Veda. 577 

a few forms belonging to each of the four pairs, and enumerated 
pp. 449-52, radical d and Osteins. 

The archaic Vedic endings -d, -dsas, -d, and -ebhis become 
entirely obsolete in classical Sanskrit. We may therefore expect 
d priori that those parts of the Veda itself which are character- 
ized by the greatest frequency of these endings belong to the 
oldest periods of Vedic poetry ; and that those in which they are 
least frequent approach the later Sanskrit most nearly in time. 

In order to see whether this theory is borne out by facts, let 
us apply our criterion in comparing 

the Riksanhita with the Atharvasanhita — §2 ; and 
different parts of the Riksanhita — §§4— V. 

§2. The Atharvasanhita, as is perfectly well known, is of much 
later date than the bulk of the Riksanhita. This is proved by 
the internal evidence of its contents, and by many other criteria 
of recognized authority in literary history. 

The truth of this fact is confirmed in the most unequivocal and 
striking way by comparative statistics. The differences between 
the two sanhitas are large and are all in the same direction. 

The following table shows, first, the actual number of occur- 
rences of each ending in the Rik, and then in the Atharvan. 
Next it gives the ratio of the number of occurrences of each 
archaic form to that of its modern equivalent for the Rik, and 
then for the Atharvan. Finally, it compares the ratio of each 
pair in the Rik with the ratio of the same pair in the Atharvan. 



RV. 
AV. 


-d 

2391 

102 


-au 
293 
474 


■dsas 
1037 

57 


■ds 

2180 
1366 


-d 
1682 

228 


-dni 

1050 

334 


-ebhis 

571 

43 


-ait 

6G6 
226 


RV. 
AV. 


816 : 
22 ; 


: 100 
: 100 


48 : 
4 : 


: 100 
: 100 


160 : 

68 : 


100 
: 100 


86 : 
19 : 


; 100 
100 



37 : 1 12 : 1 2* : 1 4} : 1 

For example, the Rik has 2391 duals in -d against 293 in -au, 
while the Atharvan has but 102 in -d against 474 in -au; or, the 
RV. has 816 archaic forms to 100 modern forms, while the AV. 
has but 22 archaic forms to 100 modern. That is, the d-forms 
when compared with the aw-forms are 37 times more frequent in 
the old text than in the young. In other words, the archaic 
forms are rapidly going out of use in the later Vedic period. 
The evidence of the other pairs all points in the same direction. 
While the RV. has 48 forms in -dsas to 100 in -ds, the AV. has 
but 4 ; or, the old form is 1 2 times more frequent in the old text. 
In like manner, the old form of the instr. is 4J times more frequent 
in the old text than in the young ; and that of the pi. n., 2£ times. 
Cf. also pp. 342-9. 

§3. The period of the composition of the hymns belonging to 
the original Rik collection was a long one: this is a fact well 
established by internal evidence, e. g., by the change in the 
relative positions of Varuna and Indra. Moreover, it is no longer 
disputed that there was also a wide interval of time between the 



578 C. R. Lawman, [Add. note 2. 

poets of the Veda and the diaskeuasts. But there are hymns (as, 
iii.29: v.87: vi.46; 47: vii.33) which were added even after the 
close of the original collection, and the date of some of which is 
subsequent to that of the diaskeuasts. 

If, now, the difference in time between the Rik and the Athar- 
van is so plainly shown by comparative statistics, why may not 
the different parts of the Rik be chronologically arranged, or at 
least separated, by the same method ? 

The answer to this question depends upon the way in which 
the Rik text is divided into " different parts." a. The most 
natural division is that into mandalas ; see §§4-6. b. The one 
most fruitful in striking results would be the division into groups 
whose temporal relations are well known by other criteria — con- 
tents, vocabulary, etc. ; see §7. 

§4. Comparison of the different mandalas. For the purpose of 
this comparison, I have made tables, the most significant of which 
follow. 

The different mandalas are of very unequal extent. The evenly-printed pada- 
text of Miiller's second edition fills about 842-j- pages. Line 1 of Table a gives 
the number of pages actually covered by each book in Miiller's text; and line 3, 
the number of pages which each book would occupy, if the whole were evenly 
printed or written on 1000 pages. Line 4 gives the rounded percentage of text 

Table a. 
73-j- 110 
vii. viii. 
87 131 
.09 .13 

From this it appears that the family books ii.— viii. are arranged according to size 
in ascending series. The books ix., x., and i. form again another ascending series; 
but if book i. were divided into its 14 family collections, these would all come 
before book ii. 

Table 1 gives the actual number of occurrences of each old 
form and of each new form in each mandala of the Rigveda. 
Thus in the first mandala, the ending -dsas occurs 187 times; and 
the ending -ds, 412 times. The forms of the Valakhilya are 
counted with those of book viii. A form or two may have been 
counted twice (so jajndnd's, x.14.2, under Jaw and jUd), and some 
duplicates in the same verse (not given by Gr.) omitted. 

Table 2 gives the percentage of the old form of each pair as 
compared with the sum of the old and new of that pair in each 
mandala. By implication, it gives the percentage of each new 
form also. Thus in book ii. there is occasion for using the dual 
in -d or -au in (92+21=) 113 instances. In 81 4 per cent of the 
whole number (113 : 92=1.00 : .81), -d is used; and in the rest, 
(100— 81 4 =) 18 6 per cent, -au is used. In this way the absolute 
size of any mandala may be disregarded. The last column gives 
the percentage for all ten mandalas. Thus 2391+293=2684; and 
2684: 2391=1.00: .89 1 . 

Table 3 gives the mandalas arranged according to the relative 
frequency of each of the four archaic endings, or (what is the 
same thing) according to the relative infrequency of each of the 



for each book. 






39 51 48i 


60 


62£ 


ii. iii. iv. 


v. 


vi. 


46 60 58 


71 


74 


.05 .06 .06 


.07 


.07 



76 


153 


169 


842* 


ix. 


X. 


i. 


Sum 


90 


182 


201 


1000 


09 


.18 


.20 


1.00 



Add. note 2.] Noun-Inflection in the Veda. 



579 



four modern endings. It is based on the percentages of Table 2. 
Thus book viii. has the maximum percentage, 41, of the archaic 
-dsas; and, of course, the minimum, (100 — 41=) 59, of the modern 
-as. The last column shows the difference between the maximum 
and the minimum for each set. This difference is largest for the 
fourth set ; but there the aggregates of Table 1 are small. 

Tablk 4 gives the most numerous mid important agreements 
between the four series of Table 3. Wherever the sequences of 
any two series are contradictory, the spaces for the books con- 
cerned are left blank. What results, now, may be deduced from 
these arrangements ? 

Table 1. 





ii. 


iii. 


iv. 


v. 


vi. 


vii. 


viii. 


ix. 


X. 


i. 


Sum 


-d 


92 


HO 


122 


214 


196 


246 


449 


35 


315 


612 


2391 


-au 


21 


10 


7 


40 


20 


25 


32 


3 


62 


73 


293 


-dsas 


47 


66 


61 


83 


70 


113 


142 


103 


165 


187 


1037 


-as 


95 


143 


134 


no 


143 


219 


204 


198 


462 


412 


2180 


-d 


105 


101 


118 


104 


150 


145 


207 


166 


272 


314 


1682 


-dni 


56 


11 


63 


49 


96 


99 


115 


112 


176 


213 


1050 


■ebhis 


25 


52 


25 


31 


61 


53 


76 


33 


92 


123 


571 


-ais 


32 


53 


47 


48 


69 
Table 


63 
2. 


69 


23 


121 


141 


666 




ii. 


iii. 


iv. 


v. 


vi. 


vii. 


viii. 


ix. 


X. 


i. 




■d, d. 


.81 4 


.91' 


.94 6 


.84 3 


.90' 


.90 s 


.93 3 


.92' 


.83« 


,89 3 


.89' 


-dsas 


.33 1 


.31" 


.31 3 


,32 8 


.32 9 


.34° 


.41° 


.34* 


.26 3 


.31* 


.32 2 


-<!, p. 


.65" 


.58' 


.65" 


.68° 


.61° 


.59 4 


.64 3 


.59' 


.60' 


.59 6 


,61 6 


-ebhis 


.43 9 


.49 M 


.34' 


.39 s 


.46 9 

Table 


.45' 
3. 


.52 4 


.58" 


43 2 


.46 6 


.46 9 


I. 


iv. 


viii. 


ix. 


iii. 


vii. 


vi. 


i. 


v. 


X. 


ii. 


Diff. 


-4, A. 


.95 


.93 


.92 


.92 


.91 


.91 


.89 


.84 


.84 


.81 


.13* 


II. 


viii. 


ix. 


vii. 


ii. 


vi. 


v. 


iii. 


iv. 


i. 


X. 




•dsas 


.41 


.34 


.34 


.33 


.33 


.33 


.32 


.31 


.31 


.26 


.14' 


III. 


v. 


ii. 


iv. 


viii. 


vi. 


X. 


ix. 


i. 


vii. 


iii. 




< P- 


.68 


.65 


.65 


.64 


.61 


.61 


.60 


.60 


.59 


.59 


.09 3 


rv. 


ix. 


viii. 


iii. 


vi. 


i. 


vii. 


ii. 


X. 


v. 


iv. 




•ebhis 


.59 


.52 


.50 


.47 


.47 


.46 


.44 


.43 


.39 


.35 


.24 2 












Table 4. 














iv. 


viii. 


ix. 


iii. 


vii. 


ii. 


vi. 


i. 


v. 


X. 




-d, d. 


.94 6 


.93 3 


.92' 


.91' 


.90 s 




.90' 


.89 s 


.84 3 


.83 8 




•dsas 




.41° 


.34 2 




.34° 


.33' 


.32' 




.32 8 


.26 3 




-d, p. 


.65 2 


.64 3 


.59' 




.59 4 














■ebhis 




.52 4 




.49' 


.45' 


.43 9 








.43' 





.. § 5 :. * n tne fi * st . P lace > negatively, as between the family books 
ii.— vii., the statistics fail to show any clear difference in time. This 
appears (a) from Table 2, in which the differences between most 
of the books ii.-vii. for any one ending are generally too small to 
admit of any conclusion respecting a corresponding difference in 
time. Thus in books ii.-vii., the percentages for -dsas range only 
from 31 s to 34°; and for -ds, of course, from 68' to 66". Further- 
more, (b) there is too little special agreement between the four 



580 0. B. Lanman, [Add. note 2. 

series of Table 3. Thus in the series based on the frequency of 
the first archaic ending -d, book iv. has the maximum, and book 
ii. has the minimum ; while for the fourth archaic ending -ebhis, 
book iv. has the minimum ; and for the third, book ii. stands next 
to the maximum. 

Table 4, by leaving those spaces blank in which the four series 
are at variance, makes this lack of agreement evident at a glance. 
Thus according to the first and fourth series, book iii. stands 
between books viii. and vii. ; while according to the second and 
third series (Table 3), it stands after them both. 

On the contrary, as between the RV. and AT., the percentages for -a range 
from .89 (RV.) to .18 (AV.); and for -au, from .11 (RV.) to .82 (AV.);— for -dsas, 
from .32 to .04; and for -ds, from .68 to .96. That is, the differences are large 
and in the same direction. 

Positively stated, then, the result is that the family books 
ii.-vii. are in general of about the same age, so far as the evi- 
dence of forms goes. And this result is quite in consonance with 
what we should infer from the other criteria. 

§6. In the second place, positively, as between books viii., ix., 
vii., and x., a rude chronological arrangement may be made. 





Till. 


IX. 


Vll. 


X. 


I. 


.93 


.92 


.91 


.84 


II. 


.41 


.34 


.34 


.26 


III. 


.64 


.60 


.59 


[.61] 


IV. 


.52 


[.59] 


.46 


.43 



Here we see a complete agreement of all four series in every 
case but two. Taking the seventh as a fair representative of the 
family books ii.-vii., our result indicates that the eighth is older 
than the other family books ; and that the tenth is youngest of all. 

Now the tenth is, on the average, confessedly the latest of all, 
and so far our result is correct. 

The eighth, or book of the family of Kanva, is distinguished 
from the other family books by very essential characteristics. 
Its redaction is by the hand of an entirely different diaskeuast ; 
its metres are in general quite different ; the responsive or dra- 
matic structure within the strophe (e. g. viii.50), and the concate- 
nation of the strophes themselves (e. g. viii. 18), striking and 
peculiar. Table 3 also distinguishes it quite sharply from the 
other family books : in series II. and IV. it precedes all of them ; 
in series I., all but book iv. ; and in series III., all but v., ii., and 
iv. While, then, these figures indicate that book viii. is older 
than the other family books, I will not lay stress on this result 
until the relations of book viii. to the rest have been more care- 
fully determined by other criteria. 

The Soma-book falls between viii. and vii. according to our table. 
This result is acceptable enough, especially as it places book ix. 
decidedly before x. ; for we can hardly say that the book is not 
antique, despite its wearisome character. But its tabulated dis- 
tances from viii. and vii. are too small to have much weight. 



Add. note 2.] Noun- Inflection in the Veda. 581 

The Atharvan is much younger than the Eik. But the Atharvan has some 
passages of decidedly antique cast; and the Rik has passages of as modern 
stamp as the average Atharvan hymn. In just the same way, the tenth mandala 
is, on the average, the youngest book of the Rik ; and the seventh, on the average, 
one of the oldest. But the tenth has unquestionably some parts as old as any- 
thing in the whole collection ; while in the seventh occur passages which date 
from the latest epoch of Vedic poetry. 

After all, then, we could hardly expect that a comparison of different mandalas, 
whose periods of origin overlap and interlace, would yield figures susceptible of a 
rigorous interpretation. Besides, single hymns, like i.116; 117: ii.39: x.106, 
sometimes disturb the numerical relations inordinately. 

§7. Far more striking, I am convinced, would be the results if 
we tried this same criterion upon groups of hymns marked as 
especially old or especially young by their contents, vocabulary, 
phraseology, metre, place in the collection, etc. It would not be 
difficult to pick out a mass of text (made up of such late hymns 
asi.133; 162; 164; 191: iii.29: v.44 : vii.33;50: x.85; 90; 109, 
and perhaps the last thirty of the book, and of the frequent addi- 
tions at the end of hymns like viii. 4.19-21) in which the compar- 
ative absence of the archaic forms -d, -dsas, -a, and -ebhis would 
reflect a striking light on the chronology of the parts selected. 
This evidence of their youth might then be confirmed by noting 
the absence of other antique forms (such as homophonous instru- 
mentals, locatives in -d, case-forms without ending, etc.), and the 
presence of modern forms (such as the forms of transition from 
the I to the ^declension). The like method might also be applied 
with good results to some of the groups of the first mandala. But 
since my collections do not include the material necessary for this 
investigation, I will leave the subject here. 

3. Numerical Summary of the Noun-forms in the Riksanhita. 

§1. The forms in the Rigveda which exhibit distinctively nom- 
inal inflection are more than 96,000 in number. Table 1 accounts 
for 93,277 forms, and the following paragraph for 2727 ; these two 
sums amount to 96,004. 

Forms from the stem sdkhdy (192), from stems in ai (259), o (750)J au (40), t and 
d (14), th (90), m (112), v (647), saradbhias and bandhupr'chd (2) were not included 
in Table 1 because it did not seem worth while to increase its size by so many 
new columns; vifvam (139), yena (57), Una (97), and forms in -end (85) were 
excluded because it was not worth the trouble to separate them into masculines 
and neuters; likewise vifvds (104) and mai-utas (133), which were not separated 
into nominatives and accusatives. A few doubtful forms were excluded here 
and there — so half a dozen genitives plural in -dm, p. 353. 

The number of excluded forms here noted is 2727. All other uncounted noun- 
forms are, I think, quite insignificant in number. The exclusion of these forms 
(most of them nom. or ace.) from Table 1 does not materially affect the truth of 
the statements based upon it. Forms showing distinctively pronominal inflection 
are, of course, not included : cf. p. 325. 

§2. Table 1 shows at a glance the general distribution of all 
the forms. It gives for each class of stems the actual number of 
occurrences of each case : first the masculines— singular, dual, and 
plural ; then the femiuines — s., d., and p. ; and last the neuters — 
s., d., and p. 

vol. x. 77 



582 0. R. Lawman, [Add. note 3. 

The okder of the stems coincides with the order of their treatment in the 
body of this article. Grouping them as on page 321, we see that columns 1-8 of 
Table 1 contain the stems whose suffixes end in vowels — group I. ; columns 9-24, 
the suffixless stems — group II. ; columns 25-38, the stems whose suffixes end in 
t, n, and s — group III. 

For example, the number of vocatives singular masculine from stems in I will 
be found in the second column, eighth line; it is 1115, i. e. 1105 regular forms 
(see p. 389) and 10 transition-forms (p. 390 top). 

The fourth line from below gives the sum of forms in each stem-class for all 
cases together. The last three lines give the sums for each gender. Thus there 
are in all 9526 occurrences of forms from stems in I: and of these, 6012 are 
masc. ; 3111 are fem. ; and 343 are neuter. 

The thirty-ninth or right-hand column gives the sum of forms for each case 
from all stems together. Thus there are in all 2442 datives singular masculine. 

§3. Noun-forms. The noun-forms of the Rik are more than 
96,000. The personal verb-forms number only 22,461, according 
to Prof. Avery's count in the Journ. Am. Or. Soc'y, x.319. The 
former are therefore between four and five times as numerous as 
the latter. 

§4. Stem-finals. The fourth line from the bottom of Table 1 
shows that the forms from stems in d, and d are vastly more 
frequent than those of any other stems. They number 46,836 or 
about one half of the sum total : i. e., they are nearly as numerous 
as those of all the other stems put together. 

Then follow in order the stems ending in \, with 9526 forms 
then those in #, with 5670; in as, with 5221 ; % (B), with 2895 
ant, with 2750; r, with 2553; man, with 1778; in, with 1271 
an, with 1240; vant, with 1089. For all other stem-classes, the 
totals fall below 1000. 

By adding the numbers of the last line in columns 1-8, we 
obtain the number of neuter forms (12,050) from stems of group! 
whose suffixes end in vowels. Performing a similar addition for 
each group and each gender, we get the following figures. 

Table 2 

I. Col's 1-8. II. Col's 9-24. III. Col's 25-38. Sum 

Masculines 44,779 2939 10,232 57,950 

Peminines 11,980 3327 602 15,909 

Neuters 12,050 849 6519 19,418 



68,809 7115 17,353 93,277 

From this it appears that 68,809 forms, or about three-quarters 
of all, come from stems of group L ; 7115 from (II.) suffixless 
stems'; and 17,353 from (III.) stems whose suffixes end in t, n, or s. 

Radical themes in i and A, though belonging to group II, are included in I. In 
certain cases the fem. and neuter forms of stems in radical d and a are not dis- 
tinguishable from those in suffixal d and a (pp. 439 med., 440), so that groups I. 
and II. cannot be entirely separated. And so the figures just given are not 
strictly accurate. 

§5. Genoers. Table 2 shows further that the masculine forms, 
57 950 in number, are more frequent than the feminines and neuters 
together. These are not so widely different in frequency: the 
neuters come second, with 19,418 forms; and the feminines last, 
with 15,909. 



NUMERICAL SUMMARY OF THE NOUN-FORMS IN THE RIKSANHI 









a, d 


i, A 


t,B 


t, C 


u, A 


B 


■&,C 


r 


d 


& 


ac 


c 


j> 


* 


d 


dh 


n 


P 


bh 


n 


f 


sh 


* 


h 


/ 


ant 


at 


vas 




r 


N. 


10071 


1498 


5 


70 


905 




26 


870 


273 


24 


47 


4 


47 


30 


98 


6 


105 




2 


8 


20 


2 


7 


48 


79 


836 


178 


240 






A. 


6861 


1262 


1 


31 


740 


.... 


28 


225 


115 


26 


33 


24 


61 


51 


58 


27 


17 


1 


3 


26 


43 


5 


17 


69 


46 


325 


14 


67 




g 


I. 


411 


68 


1 


3 


172 


._ 


4 


8 





1 


1 


4 


39 


11 


18 


1 


1 


1 


3 


. - „ . 


2 


1 


2 


3 


4 


32 


2 


5 




D. 


1107 


166 




8 


247 




6 


58 


9 


7 


3 


3 


9 


4 


7 


4 


7 





2 


2 




9 


17 


47 


30 


226 


9 


146 




Ab. 


183 


15 






34 






16 


__ 


1 


1 





2 


1 


1 









1 









1 


9 


3 


25 


1 


3 




N 


G. 


1890 


223 


3 


8 


245 




2 


132 


2 


5 


3 


2 


4 


6 


3 


2 


4 


1 


1 


3 




1 


9 


67 


8 


180 


21 


63 




L. 

V. 

A.V. 


1235 

2498 
1316 


143 

1115 

340 


~~"l 


13 


38 
508 
201 




21 


9 
110 

148 


21 
34 


12 
3 


1 
~~8 


2 
1 
4 


1 
4 

47 


22 
10 


3 

1 

14 


~17 


9 
36 

8 








6 


— 


3 
1 
6 


""i 

10 


1 
15 
17 


13 




28 
14 










w 


2 




5 


92 


10 


'A 


J 
■« 


I. 

D. 
4b 


8 
7 
4 


24 

7 






8 
1 
2 
















































CO 






3 




































3 






















2 






















P 


G. 
L. 

N.V. 


27 

27 

3217 


5 

25 

541 


2 


1 

"58 


2 

21 

746 




41 


14 

6 

295 










1 


























1 


1 


--- 


84 


11 


1 
22 






1 
64 






















41 


63 


208 


32 


4 


13 


4 


26 


27 


14 


12 


31 


49 


433 


53 


44 






A. 


908 


91 





13 


196 


.. 


2 


65 





7 


22 


8 


12 


7 


9 


14 




3 




9 


7 


6 


14 


23 


12 


89 


3 


14 




►4 


I. 


871 


258 





4 


210 


__ 


2 


112 


3 


15 






9 


35 


8 


„ - _ 


1 






1 


1 




1 




2 


33 


9 


3 




Cm 


D. 

Ab. 

G. 

L. 


130 
25 

477 
548 


44 

17 

128 

37 


— - 


1 


18 

5 

47 

17 


-- 


3 

2 


70 
4 

62 
4 


1 

" 2 


1 






1 


13 

12 

40 

8 






















2 


11 
1 

60 
6 


5 


5 


























1 


5 


1 

2 






5 


2 
1 


3 


1 
3 


3 


- 


1 
1 






4 


3 
1 


3 
1 











r n. 

A. 
« I. 

1 D - 
§ Ab. 

* s 

v. 

N.A.V. 
I. 

g Ab. 

P G. 

L. 

N.V. 

A. 

J I. 

a d. 

5 Ab. 

fi G. 

L. 


1053 

522 

656 

20 

8 

54 

30 

95 

321 

2 


474 

609 

183 

551 

20 

45 

256 

29 

23 

1 


734 

303 

82 

31 

28 

116 

56 

117 

372 

1 

1 

21 

8 

435 

312 

113 

3 

4 

73 

65 


50 

87 

111 

46 

16 

3 

14 

1 


73 

51 

12 

13 

6 

28 

1 

8 

12 


1 
1 
1 

1 
1 


26 
64 
42 
18 
5 
21 
14 

15 


103 

36 

1 

3 

10 

33 

4 

20 

38 


18 
81 
10 

3 

8 


16 

52 
9 


— 


12 
91 
44 
4 
3 
26 
20 


8 
61 
12 
13 

1 
24 

1 


20 
25 
36 

3 
49 

3 
29 


10 
15 
13 
11 

14 

5 


4 

15 

46 

36 

3 

1 

7 


"20 

7 


4 
14 

4 

12 

5 


2 
12 

5 
17 

21 


28 

38 

74 

3 

2 

1 

15 


13 

17 

15 

13 

3 

7 
12 


79 
26 
20 
20 
24 
9 


2 
4 
9 

2 


3 
10 
8 
4 
5 
6 


4 
5 
2 


:::: 


... 


... 




2 
2 










5 
1 
3 
3 


... 


















1 
3 










m 


... 


6 


1 


6 


3 


5 








12 


2 


... 


1 








.... 


£ 
































































1 

4 

3 

606 

393 

295 

8 

18 

47 

182 


2 

2 

1 

300 

126 

376 

9 

5 

98 

61 




















































3 

3 

61 

83 

44 

1 

32 
17 








1 

4 
48 
9 
8 
1 
1 
4 
4 










2 

2 

13 

10 


1 


















1 


... 


2 








in 


1 

91 

8 

7 

4 

10 

3 


-- 


17 
53 
39 
13 
1 

19 
36 


28 

~29 
4 


1 

_ 27 


— 


21 
9 
1 


























49 

49 

1 


14 

34 

2 


23 

79 

3 

1 


--- 


170 

175 

15 


6 

8 


77 

155 

91 


49 

46 

2 


23 
125 


9 
4 


11 
10 


24 
18 
11 






















































12 

117 
81 
























~~7 


1 


... 


1 


9 
1 


2 
34 


2 
32 


5 

7 


... 


1 


13 
13 


43 

48 


9 






2 


.... 


... 


... 



N. 



N. 



N.A. 

I. 

D. 
Ab. 

G. 

L. 

V. 
A.V. 

I 

D. 
Ab. 

G. 

L. 
A.V. 

I. 

D. 
Ab. 

G. 

L. 



Sum 
Masc. 
Fem. 
Neut. 



4275 

421 

395 

206 

695 

1256 

3 

57 

3 


256 






428 

56 

6 

11 

168 
49 


-- 


15 7 


5 


78 
3 

10 

10 
3 

40 


159 
5 


1 


3 
1 


6 
2 


15 
22 
37 
6 
5 
13 










151 
1 
1 


5 
2 


"2 


35 
31 


1 

4 

33 

11 

41 


12 


272 
46 
14 

6 
23 

5 


18 
1 

30 
1 


2 

2 

"i 






1 


.... 


3 


1 






1 










- ~ 




6 

1 


... 


1 




















1 


16 

1 


.... 


1 




1 


... 


1 








20 
5 


1 








































5 






3 


-■ 








1 


1 


























2 


... 


3 






































































1... 


















2 
3 
























































































1 




















































1 


















2732 

338 

1 

11 

50 

249 


63 






204 
6 










31 
18 
































6 

7 






















2 










2 


























































































1 














44 
i 










6 

434 
116 
118 
200 










9 

554 
290 
155 
109 










1 

785 

82 

522 

181 










1 
1 

2750 
2366 


356 
306 


637 
632 


1 












737 

544 

188 

5 


308 
142 

166 


332 

93 

238 


468 
305 
158 


773 
458 
307 

8 


345 
106 

235 
4 


224 

196 

27 

1 


633 

24 

609 


88 
16 
72 


401 

123 

270 

8 


375 

38 

335 

2 


193 
94 
32 
67 


465 

312 

61 

92 


357 

272 

72 

13 


46836 

31821 

4318 

10697 


9526 

6012 

3171 

343 


2895 
20 

2875 


788 

215 

572 

1 


5670 

4363 

328 

979 


5 
5 


536 

137 

383 

16 


2553 

2211 

328 

14 


1 5 


384 


50 


5 



[SANHITA. TABLE 1. 



t 


at 


«as 


^os 


t>an£ 


mant 


van 


man 


an 


in 


as 


is 


us 


Sum 


6 
5 
2 
6 
5 

3 

2 


178 

14 

2 

9 

1 

21 

10 


240 

67 

5 

146 

3 

63 

28 
14 


56 
7 
4 
7 
6 
7 
2 
2 


192 
86 
25 
51 
3 
88 
18 

162 
22 


100 
93 
14 
13 

24 
8 
8 
5 


308 
68 
12 
32 
2 
38 
13 

179 
34 


207 

100 

96 

22 

6 

18 

14 

24 

6 


376 

134 

9 

48 

2 

57 

24 

105 

103 

2 

1 


213 

192 

13 

34 

4 

71 

3 

76 

401 

4 

2 


346 

252 

36 

74 

13 

60 

8 

190 

119 


8 

22 

"ii 

"8 
2 


10 
6 

5 
10 

4 
44 

"l 

4 


17315 

11136 

1012 

2442 

338 

3315 

1563 

5112 

3044 

46 


3 






















24 










1 
3 














9 


1 


1 








1 


.... 

35 
2 

16 
3 


2 

81 

13 

20 

4 


17 

1 

153 

44 

23 








75 










83 


3 
9 
3 
1 


53 
3 
9 


44 

14 

3 


1 

22 


91 
54 
19 
L9 


53 
16 

7 
2 


77 

33 

48 

3 


287 
51 
12 

7 


24 
1 


14 
6 


6951 

1776 

1723 

329 


1 








65 



6 


5 


5 


... 


14 
11 


1 


31 
4 


1 


2 
1 


7 
2 


20 


... 


6 


940 
652 













3 
1 


2 
4 


2 
1 





86 

60 

30 

4 

5 

45 

5 

61 

29 


1 


4 














































































































2 
1 






... 


1 


_.... 




... 


.... 


----- 


1 


1 


.... 
















































































































9 
1 


.... 


95 

41 

1 


... 


1 














1 
2 












1 














































1 


1 






27 





















2751 

2195 

1504 

827 

192 

499 

473 

339 

870 

4 

1 

4 

37 

40 

2215 

1762 

1016 

28 

44 

517 

591 



2 
6 
4 


18 
1 


2 
2 


82 

18 

13 

1 

2 


204 
10 

4 


81 
5 


9 
5 

33 
1 
1 

11 


430 

142 

32 

14 

73 

209 


49 
6 
11 
25 
59 


7 
3 

1 


1421 

710 

225 

107 

304 

45 

2 

17 


320 

141 

2 

2 

28 

50 

2 


143 

31 

11 

5 

7 
5 

1 


8441 

1718 

823 


R 


30 
1 


1 


393 


3 

5 


6 
3 


1 


1460 
1753 

7 


3 














5 


7 


.... 


102 
















3 
























































2 
















1 












4 




























6 

7 






1 


1 

2 


2 
1 
1 


15 
4 


223 
67 
2 
1 
6 
9 


18 

34 

1 

1 

28 

3 


.... 


284 

127 

1 

1 

3 

10 


26 
12 


23 
1 


3629 
621 










6 












2 
1 


4 


15 


1 














138 


1 












9 


303 




356 
306 


637 
632 


231 
114 


1089 

859 







6 


435 
344 


978 
883 

7 


1778 

551 

13 


1240 

984 

14 

242 


1271 
1260 

11 


5221 

1475 

489 

3257 


663 

76 

1 

586 


347 

110 

6 

231 


93277 
57950 
15909 


4 


50 


5 


117 


230 


j)i 


88 1214 


19418 



Add. note 3.] Noun- Inflection in the Veda. 583 

About half of all the feminines come from stems in 4 (4318) 
and i (3171) — see Table 1. last line but one. In group II., the 
feminines (3327, root-words) almost equal the sum of masculines 
and neuters — see Table 2. Stems in I (p. 367) serve as feminines 
to most of the stems of group III. The feminine forms from 
adjective as-stems and ushds number 489. The feminines from 
stems in vocalic root-(-« number 72 ; but these have a certain 
right (p. 327) to be in group II. Aside from these 561 forms, the 
feminines in group III. are so very few — only 41 — as to be fairly 
called sporadic. Nevertheless, it is instructive to observe that of 
all the feminines, barely one-fifth (2875) are from stems in deriv- 
ative i. 

The paucity of the neuters (849) in group II. is also worthy of 
notice. The neuters of this group consist chiefly of adverbial 
accusatives in -ak, -k, forms from " radical <i-stems," from hr'd, 
-pad, d's, bhd's, and mdh, dhar, 'd'dhar, and s&ar. 

The fewness of datives in the neuter gender (s., 823 ; d., ; p., 6) 
when compared with the datives masculine and feminine (3651) 
is worthy of remark, as showing that the dative relation most 
often implies a personal subject : i. e. one gifted with sex. 

§6. Numbees. Table 3 shows that the singular forms amount 
to 65,608, or more than two-thirds of the grand total. The 
plurals number 23,321 ; the duals, 4348. I doubt whether any 
Indo-European language that had a dual number would show in 
any recorded text so large a proportion of duals as does the 
Rigveda. 

Table 3. 

Singular Dual 

Masculines 42,233 3281 

Feminines 8780 956 

Neuters 14,595 111 

65,608 4348 23,321 93,277 

§7. Cases. With the nominative coincide : the vocative (apart 
from the accent), in the dual and plural of all genders ; the accu- 
sative in the dual of all genders, in the singular and plural of the 
neuter, and often in the plural of the feminine. For these cases, 
therefore, the forms have not been separated when coincident. 
For the remaining five cases, however, the forms, even when 
coincident, have always been separated — usually by Grassmann's 
exegesis of the passages in which they occur. 

Table 4. 

N.A.T. Inst. Gen. Loo. Dat. AM. 

Sg.m. + f. + n. 47,296 4234 5274 3789 4092 923 

Du.m. +f.+n. 4016 53 116 123 25 15 

Pl.m. + f. + n. 16.333 3360 1595 1546 363 124 

Sum total 67,645 7647 6985 5458 4480 1062 

To the first group of cases belong about three-quarters of all 
the forms : i. e., the nominatives, accusatives, and vocatives 
together number 67,645. Of these the nominatives are of course 



Plural 


Sum 


12,436 


57,950 


6173 


15,909 


4712 


19,418 



584 C. R. Lanman, [Add. note 3. 

much the most frequent, and the vocatives least frequent. Thus, 
in the masculine and feminine singular, where these cases differ 
in form, the nominatives number 20,066; the accusatives, 13,331 ; 
and fhe vocatives, 5451. 

The remaining five cases in the order of frequency as shown by 
the sums total of Table 4 are: instrumental, with 7647 occur- 
rences ; genitive, with 6985 ; locative, with 5458 ; dative, with 
4480 ; and last of all, the ablative, with 1062. 

The most frequent case of all is the N.s.m. with 17,315 occur- 
rences ; cf. Benfey's remark about the prototypical influence of the 
form of the nom.s., Vedica, p. 115. For the /(-stems alone, Table 1 
shows not a single nominative. Of the D. and L.d.n. there is not 
a single example in the Rigveda — see Table 1, right-hand column. 
Of the D.d.f. there is but one example, r6dasibhydm, i. 136.6. 

The blank spaces of Table 1 show how often it happens that, 
even for whole stem-classes, we cannot make up a complete para- 
digm of forms in actual use. As for single stems, I do not believe 
that there is one in the whole Vedic vocabulary of which a com- 
plete set of declensional forms occurs in the Rik. Devd, rdtha, 
ff6, vr'shan, and others are nearly complete. All the singular 
forms of indra occur (see p. 340 top), and of vr'ka except the 
loc. Cf. also the remarks on p. 350, D.p.n. 

In Indo-European the ablative appears to have been formally 
distinguished from the other cases only in the singular. Here 
the ablative ended perhaps in -at; the genitive, in -as. In the 
dual, the ablative was made to share the same form with the 
instrumental and dative ; and in the plural, the same form with 
the dative. Why now did the ablative dual and plural have no 
case-form of its own ? 

Upon this question the statistics of Table 4 cast a most inter- 
esting light. Undoubtedly the ablative dual and plural has no 
case-form of its own because these relations were of extremely 
infrequent occurrence. 

Thus in the whole Rik text, there are only about 124 instances in which there 
is occasion for expressing the ablative plural relation. It is therefore in perfect 
accordance with the principle of linguistic economy that there should be no 
separate form devoted exclusively as it were to the expression of this relation. 
And in fact it is made to depend for its expression upon the case-form of the 
dative, which — be it observed — is, next to the ablative, itself the most infrequent 
of all cases. 

Similarly in the dual, one case-form is made to serve as the means of expression 
for the three relations — instrumental, dative, and ablative. Here again, the three 
least frequent relations (I., 53; D., 25; Ab., 15) are united under one form ; and 
the two more frequent relations (L., 123 ; G-., 116), also under one form. 

In the singular, there is occasion to express the ablative relation 923 times. 
In 534 of these instances the word has a non-a-stem, and here the old distinction 
of form between genitive and ablative has become effaced and the -as or s of the 
genitive is made to suffice also as an expression for the ablative. But in the 
remaining 389, or nearly half of these instances, the stem ends in a and the ablative 
is clearly distinguished in form from the genitive by the ending -dt— see p. 337. 
The original state of things, therefore, is very fairly reflected in the language of 
the Veda, although the Avesta has retained the ablative ending in the singular of 
non-a-stems. 



Corrections etc.] Noun-Inflection in the Veda. 585 

CORRECTIONS AND ADDITIONS. 

p. 327, 11. 19, 20. Read the few masculines and the feminities in t and ■&. 

p. 328, 1. 15. Read yakrt. 

pp. 334-5. As instrumental may be taken further: ddcd', i.127.7 (of. p. 
490d) ; vicvd, 'allenthalben,' x.20.8 ; sunirmathd, iii.29.12 ; havd, x.27.8 ; himd', 
x.37.10; md'nushd, i.51.1. 

p. 337, 1. 17 from below. Add viridyaft, vi.30.1. 

p. 339, 1. 9 from below. Adverbial retraction of accent is seen in diva, and 
perhaps in kshapas, didr'kshu, giihd, tand. 

p. 341, 1. 4 from below. For atmand, cf. p. 408c, and A. Weber, Jenaer 
Literaturzeitung, 1878; p. 82. 

p. 345, 1. 21. Add: devd's\as~\, v.51.13a; vdjds[as], vii.48.la ; cvhhrd's\as~\, 
i. 167.4a ; cucibhrdjas\as\ fem., i.79.1c. 

p. 348, 1. 6. Add: virid[ni], v.29.136. 

p. 348, 1. 8 from below. In ix.82.4, cevd may be neuter pi. 

p. 360, 1. 1. In AT. iv.5.6, asyai is G.s. 

p. 362, 1. 14 from below. Here belong perhaps gnd' hutd'sas, vi.50.15, and 
cucibhrdjds[as'], i.79.1. 

p. 363, 1. 7 from below. Cf. awezhddndonhd, p. 546d. 

p. 366, column 2 of Series B. Read i, tarn, Id', lai, Id's, id'm. 

p. '367, 1. 12. Read urmin-d. 

p. 371, 1. 13. Dele akshi.' The stem belongs to B ; p. 391, 1. 21. 

p. 375, 1. 6. Insert except in monosyllabic steins after vowel. 

p. 375, 1. 7. Read jdtya for kshaipra. 

p. 377, head-line. Read N.s.f. [vii.97.2». 

p. 381, top. For suasti, RV. v.42.46, the AT. has the modem form svastyd', 

p. 383, 1. 27. So criyai, TS. ii.2.8 9 ; meshyai, TS. iii.59. 

p. 388, 1. 13 from below. Here Delbriick puts bh&'md, vi.62.8 ; it must be 
neuter pi., pada -a, page 539 end. 

p. 389, 1. 15 from below. Add dsandi', TS. v.6.9 1 . 

p. 392, 1. 12 from below. Read divas-. 

p. 400, 1. 20. Read pdtid ; -yd, 3. 

p. 409, D.s.f. Cf. tanue, AT. i.12.4, with tanual, TS. xxiii.44. TBr. iii.7.1 4 
has tanuvai. 

p. 411, 1. 3. In i.143.3, sindhavas is G.s.f. ; cf. p. 410 end. 

p. 420. Transitions from r to a-declension : nar, nara; devar, devara ; etc. 

p. 429, 1. 25. As to mdtf'n, see "Whitney, Gr., 375. 

p. 449, T.s.m. Cf. ashdddo, Yc. xxviii.7, with rtadds. 

p. 456, 1. 7. TBr. ii.6.1 2 has prd'nk. 

p. 457, 1. 10 from below. Read occurs twice. Add viii.45.11. 

p. 462, 1. 12 from below. Read Acv. 

p. 480, 1. 9 from below. Read Locative Singular Masculine. 

p. 485. In the Br. occur kakubbhydm, anushtubbhis (Whitney, Gr., 151c). 

p. 489, 1. 24, and 490, 1. 20. Opaprk is from the root^fc. 

p. 489, end. Transition-nom. s. m., spaca-s. 

p. 490, last line. Read iii.463. 

p. 492, il 24, 23 from below. Read vdta-tvish, avidvish. 

p. 492, 1. 15 from below. Read Ted's, nds; d-fds. 

p. 501, 1. 16 from below. Read nadbhias. 

p. 504, last three lines. The four stems ought, of course, to be classed with 
the vant-stems. The N.s. would end in -an. 

p. 505, 1. 19. The form vavrdhdntas is from a verb-stem in a : Whitney, Gr., 444 

p. 512, 1. 15. Read jigivd'ft. 

p. 512, 1. 16 from below. Read c&cuvd'nsam, 6. 

p. 513. I should rather take tasthushas as Ab.s.n. : cf. Whitney, Oriental and 
Linguistic Studies, i. 138. 

p. 519, 1. 6 from below. Read 154 forms (15 stems). Add adrivas, 47. 

p. 536, 1. 24 from below. Read carman, 18. 

p. 542, 1. 13. Read several in min. 

p. 558, 1. 22 from below. An-Mhds is of course feminine. 



586 



G. R. Lanman, 



[Index of 



INDEX OF CITATIONS 

Made in the foregoing Article, from the 

RIGVEDA AND OTHER TEXTS. 

This Index is designed to include passages whose exegesis, metre, text, or 
word-forms have been discussed on the preceding pages. The nature of the 
remarks upon any particular verse is shown approximately by an Italic letter 
prefixed to the page-number. An e signifies that the remark concerns the 
exegesis of the passage ; an m, the metre ; a t, the text, its criticism, or some 
interesting variant ; a w, some word, note- worthy, modern, or irregular in form 
or otherwise. 

To facilitate reference, the distance from the head-line to the last line of each 
page io divided into four equal parts, designated by Roman a, b, c, and d after 
the page-number. 



BV. 


i. 




BV. 


i. 






BV. 


i. 




6. 


1 


ew 513a,c,d, 585. 


10. 


4 


t 386c. 




128. 


2 


w 516a, 518d. 




3 


et 339d. 




1 


e 422bc, 482c,3a. 




1 


mt 383d. 


8. 


3 


e 335a. 


11, 


3 


e 354a. 




129. 




< 469d. 


11. 


5 


w 513c. 




4 


m 515a. 






8 


m 350a. 


23. 


24 


m 338c. 




5 


m 42 6d. 




9> 


10 


m< 431b. 


24. 


14 


e 564c. 




8 


e 433a. 




130. 


3 


M> 486b. 


26. 


2 


e 562a. 


12. 


6 


et 422b. 




134. 


3 


mt 311d. 




6 


e 518d, 480bc, 


19. 


1 


emt 565d. 




135. 


9 


< 313b. 






515d. 


80. 


4 


tw 310c. 




140. 


9 


e 371b. 395c. 


27. 


2 


e 563d. 


81. 


4 


e 562b. 






13 


etfw510b; m«539d 


29. 


3 


<361a. 


85. 


3 


mt 543b; cf.318b. 


141. 


4 


e 335c. 


31. 


5 


ew 510a. 




9 


el 348b. 






8 


errew; 551d. 


32. 


8 


e 562b. 


81. 


4 


et 358d, 


495a, 


143. 


2 


to 535bc. 


36. 


1 


mt 399a. 






559b. 




146. 


4 


e 430cd. 




16 


e334d. 


91. 


18 


mw 498d, 


500c. 


149. 


4 


mtf 394d. 




11 


m 342 a. 


92. 


3 


ew 348d, cf.539d. 


150. 


1 


w 500c. 


31. 


1 


ew 526a, 330b. 




5 


e 562b. 




151. 


3 


OTf 558b. 


44. 


5 


ew 339bc. 


93. 


2 


e 562b. 




158. 


6 


< 399a. 




8 


e 482b,d. 


94. 


15 


et 386a. 




160. 


1 


344c (n. for f.). 


46. 


2 


rnw 420d. 


91. 


8 


mw 434a. 




162. 


9 


t 386c. 


50. 


6 


e 330d. 


102. 


6 


e 563c. 






16 


t 354c. 


52. 


6 


e 335bc. 


104. 


1 


e 384d, 


395b, 




20 


«408a, 416c. 


54. 


1 


e 372d-3a. 






558a. 






21 


ew 391b, 509c. 


51. 


1 


mw 382b. 




1 


mi 381d. 




163. 




£41 lb. 520b. 




2 


et 348c. 


105. 


15 


to 514d. 






10 


mt 345c. 




4 


e 312d. 


110. 


9 


to 511d. 




164. 




* 471b. 


58. 


3 


ra«o403d,419a. 


111. 


1 


mw 516c. 






15 


e 423a. 




5 


e 422b. 


112. 


10 


m* 388a. 






39 


TO* 536a. 


60. 


3 


e 356e, 348d. 




11 


et 339a. 






41 


mtf 536a. 


61. 


1 


mt 409b, cf. 414c. 




18 


etf 342d-3a 


550d, 


165. 


15 


e 552cd. 




13 


mt 338b, 514d, 






566c. 




168. 


1 


to< 558b. 






41 6d. 


113. 


11 


e 461b. 




113. 


3 


mw 539d. 


62. 


8 


m 356b. 


118. 


8 


ef 359bc. 






4 


mi 348a. 




9 


e 563d. 


120. 


8 


mtf 312c. 






1 


e 373a. 


63. 


5 


e 334d. 


121. 


1 


et 430cd. 






8 


mt 356c. 


64. 


5 


ew4l4d, 488a,cd. 




3 


mw 413b. 






9 


em 332d-3a. 


65. 


5 


t 530a. 


122. 


11 


mt 480bc, 


516a, 


114. 


7 


TO 556d, 485d 


66. 


1 


e 512cd. 






b, 519ab ; et 






446d. 


61. 


5 


m 432c, 556d,485 






509b,d. 






8 


< 348c. 






d, 446d. 


123. 


5 


et 359d. 




119. 


6 


« 341ab. 




6 


w !)00a. 


121. 


2 


to 432c. 




180. 


4 


e* 380d. 


68. 


1 


e 422b. 




6 


to 441cd. 






5 


et 373a, 376d. 


10. 


1 


t 363ab. 




1 


to 346a ; u 


490d. 


186. 


4 


* 406b. 




3 


et 353b, 423b. 




10 


mt 424bcd 




188. 


5 


w 486b. 



Citations.] 



Noun- Inflection in the Veda. 



587 



RV. 


ii. 




RV. 


iii. 




RV. 


v. 




1. 


5 


ei 518b, 519d. 


55. 


2 


344c (f. for n.). 


37. 


1 


t 558a. 


2. 


2 

4 


e 482b,d. 
m» 470ab. 




11 


i427cd. 


41. 


1 


e 449b, 42cd. 47c; 
444ab,48c, 51d, 


3. 


6 


i361a. 


RV. 


iv. 








485c. 


6. 


2 


m 334a. 


]. 


3 


et 331a. 




3 


mt 341d. 


10. 


2 


m 342a. 




19 


mt 432a. 




4 


t 345c. 




3 


to 548c. 


2. 


1 


m 524d. 




9 


t 345b. 


11. 


2 


to 350a. 




3 


e 353d. 




10 


mi 348a; e426d. 




8 


m 378d. 




11 


e 354a. 




16 


mi 45 Id. 




19 


i 333d. 




12 


ew 475b, 492a. 


42. 


4 


i 520b. 


12. 


3 


ei 447a. 




17 


w 505b. 


43. 


15 


i 386c. 


3 3. 


9 


i 38Gc. 




18 


w 516b, 521b. 


44. 




t 373d, 506a,8a. 


14. 


3 


e 416b. 


3. 


5 


to 524d. 




2 


w 508a. 


15. 


3 


e 457d-4a. 




G 


i 534a, 448a. 




5 


t 569b. 


20. 


2 


ei 381a ,443d-4a. 


16. 


20 


e 336bc. 




10 


mi 384b. 


21. 


2 


raw 498d, 500a. 


17. 


14 


ei 338d. 


45. 


6 


mt 424bc. 


23. 


12 


ew 495d. 


19. 


9 


ew 474cd. 


46. 


1 


TOi 424b-5a. 




]6 


w473d; m524d. 


21. 


6 


mt 559a, 




2 


ew 339c. 




17 


e422d; to 524d. 


23. 


1 


e 501c, 506d. 


48. 


1 


m 524d. 


24. 


11 


e 501c, 506d. 


27. 


4 


w 516b, 52] a. 


50. 


2 


to 350a. 




13 


em 349b. 


28. 


5 


m 556d, 485d, 


51. 


11 


to 526ab. 


27. 


3 


ei 508c. 






452b. 


52. 


10 


ei 339a. 


28. 


4 


mi 424bc. 


32. 


23 


i361a, 356d. 




16 


mi 43 Id. 


31. 


5 


mw 391c; e 562 


33. 


2 


ew 349b, 359d,a. 


53. 


10 


mw 515c. 






ab. 




3 


mt 529ab. 




14 


mi 420c, 42"a. 


32. 


2 


e 336b. 




5 


mt 424b-5a. 


54. 


6 


e 553d. 


33. 


2 


e 334c. 




10 


e 335d. 


58. 


1 


mw 515c. 




6 


e 379c. 


34. 


1 


e 336c. 




7 


«i412ab, 407d. 




8 


e 358c. 




9 


ew 4 1 4b. 


59. 


7 


to 393d. 


34. 


2 


e 393a. 


35. 


5 


mw 42 Od. 


61. 


16 


i 348c. 




6 


m< 548ab. 


36. 


3 


ew 414b. 


65. 


6 


mi 341d. 


37. 


4 


tw 563a. 




6 


mi 529ab. 


66. 


2 


et 348c. 


38. 


8 


ei 416c. 


37. 


4 


t 345b. 


68. 


4 


i 341ab. 


39. 


4 


et 343c. 


38. 


9 


e 448c. 


69. 


3 


et 386a. 




5 


ei381a, 573a. 


41. 


8 


e 415b. 


70. 


4 


mw 479c. 




6 


i 361a, 494ab. 


43. 


4 


m< 356c. 


73. 


5 


e 335bc. 


40. 


6 


ew526b, 527d. 




6 


« 335bc. 




8 


ei 348c. 


43. 


3 


tw 506b. 


48. 


2 


mt 371d. 


74. 


4 


w 342c. 








51. 


4 


to 548c, 550d. 


78. 


9 


to 493d. 


RV. 


iii. 




53. 


7 


w 483cd. 


83. 


3 


em 376a. 


1. 


12 


t 398d. 


56. 


5 


tow 391c, 392a. 




7 


w471b. 


4. 


1 


e 476b,cd. 


58. 




i471b. 


86. 


3 


m 538ab. 


8. 


9 


mi 345c. 




8 


ei331c. 


87. 


6 


i 386c. 


16. 


5 


tow 359b. 










9 


mt 381c. 


18. 


1 


ei 425d. 


RV. 


v. 










27. 


11 


w 48 6d. 


2. 


11 


e 332d. 


RV. 


vi. 




29. 


16 


i 513b. 


4. 


7 


m 350ab. 


3. 


1 


ew 563a. 


30. 


11 


e 373bc. 


7. 


8 


to 375d. 


4. 


1 


e 386a. 


31. 


5 


ei 457c. 




10 


ei 430cd. 




5 


t 376b. 




9 


to 493d. 


9. 


5 


TO 423b, 426d. 


5. 


5 


e 564c. 


32. 


5 


t 363a, 348c. 


19. 


4 


« 485a. 


6. 


4 


to 556d, 485d, 




9 


to 493d. 


27. 


6 


mt 535a. 






446d. 


33. 


13 


tw 341 ab, 565a. 


29. 


15 


i 348c. 


8. 


2 


to 535bc. 


34. 


3 


m< 554c. 


30. 


4 


to 432a. 


11. 


3 


e 354a. 


45. 


2 


w 462c. 




13 


£ 386c. 


12. 


4 


e 426d. 


49. 


2 


mi 406d. 57 Id. 




14 


w 355d. 




6 


mw 47 1 be. 


51. 


7 


e 389a. 


32. 


11 


e 362c. 


13. 


1 


ei 380d. 


54. 


4 


tow 391c. 


34. 


8 


i 341 ab. 


16. 


40 


e378b; ei543ab. 




19 


ei 386a. 


35. 


5 


e 336d. 




46 


t 380c. 


55. 


1 


t 546b. 


36. 


6 


i 341 ab. 


18. 


10 


et 380d. 



588 



C. R. Lanman, 



[Index of 



RV. 


vi. 




RT. 


vii. 




RV. 


viii 




22. 


6 


mt 348a. 


28. 


3 


mt 429b. 


32. 


6 


ero 446d. 


24. 


6 


t 353c, 501a. 




4 


e 362bc. 


33. 


6 


t 542c. 




7 


tw 493d. 


35. 


13 


t 445b. 




8 


e 330d. 335ab. 


26. 


2 


mt 354d. 




14 


t 345bc. 




19 


t 341ab. 


27. 


8 


tw 52 Id. 


39. 


3 


< 412c. 


38. 


2 


t 545d. 


29. 


2 


i 348c, 362a. 


40. 


4 


ew 526b, 527d. 


39. 


2 


mt 371d; e 562 


30. 


1 


m 514d. 




5 


« 552c, 362b,d. 






ab. 


33. 


4 


mi 388a. 


41. 


4 


* 389a. 


41. 


10 


to 333d. 


44. 


1 


t 520a. 


41. 


1 


e 466b. 


45. 


13 


w 462c. 




9 


m< 387d. 


48. 


1 


ew414b; e*458cd. 


46. 


12 


t 400c. 




12 


ew 523c, 538d. 




2 


ew416d, 414b. 




23 


tow 382b, 384d. 


46. 


1 


« 388d. 




3 


«363b; ew495d; 


41. 


6 


ew 502d. 




4 


et 329d. 






mt 529ab. 


50. 


16 


mt 395d. 




8 


£ 386c. 


50. 




t 341ab. 


51. 


2 


mt 462d. 


41. 


9 


w 509a. 




1 


m 342a. 


52. 


12 


t 386c. 




13 


t 386c. 




4 


toS 373b, 379b, 


55. 


9 


to 333d. 




16 


e 353d. 






394a. 


59. 


2 


e 423a. 


48. 


11 


e 562ab. 


56. 


3 


mt 546c, 345c. 


61. 


13 


e 360b. 




18 


w 513ab. 


58. 


1 


m 524d. 


63. 


1 


e 562b. 


49. 


3 


e« 487d. 




2 


w 571c. 




4 


* 529d. 




6 


et 422ab. 




6 


to* 356c. 


65. 


12 


mt 408a. 


50. 


11 


em* 516b. 


60. 


3 


t 406a, 415c. 


66. 


10 


w 51 2d. 


51. 


2 


mt 354a. 




10 


w 542d. 


68. 


3 


to 351c. 


53. 


6 


w 473d. 


61. 


2 


ew 502c, 503d. 


15. 


1 


e 336b. 


59. 


5 


« 341ab. 




5 


« 362a. 


80. 


5 


w 481c. 


60. 


14 


t 341ab. 


67. 


5 


TO 342a. 


85. 


19 


et 348b. 


62. 


8 


ew 539d; cf.388d. 


68. 


7 


ew 529d. 


90. 


10 


to 350b. 


63. 


1 
9 


m<341d, 342d. 
e S62c. 




8 


to* 379b, 380b, 
394a. 


92. 


8 


to 524d. 




10 


w 342c. 


10. 


1 


mt 341d. 


RV. 


ix. 




64. 


2 


e 568c. 


76. 


3 


et 339a. 


1. 


6 


e 518d, 480bc, 


66. 


3 


m< 356cd. 


77. 


1 


TO* 554c. 






515d. 




4 


et 358d. 


19. 


1 


< 363ab. 


9. 


4 


mt 373b. 




5 


e< 358d, 359c. 


86. 


3 


ew 405d. 


10. 


9 


e 349a. 




11 


et 393a. 




4 


em 329d. 


12. 


3 


e 373b. 389c. 


67. 


1 


to 342a. 




5 


TO 524d. 




1 


et 364d. 




8 


to< 342d. 


89. 


5 


< 380a. 


22. 


5 


to* 55 Sb. 




10 


m 346a. 


92. 


4 


e< 414c. 


46. 


4 


t 369a. 


68. 


10 


< 342b. 


96. 


1 


t 357a, 356d. 


41. 


4 


e 423a. 


72. 


2 


m 432c. 


91. 


6 


to* 345c. 


51. 


3 


et 339ab. 


74. 


4 


« 341ab. 


103. 


2 


e 389c. 


58. 


3 


e344b. 


75. 


4 


e<331c. 


104. 


13 


2 386c. 


61. 


10 


e* 376b. 




11 


tw 505d. 




18 


mw 541a. 


62. 


15 


w 385d. 




15 


< 403b. 


RV. 


viii 




65. 


12 

27 


2 358d. 
« 358d. 


RV. 


vii. 




2. 


8 


to* 536a. 




30 


tw 404c, 408c. 


1. 


19 


m» 359b. 




30 


TO*w 348b. 


68. 


10 


to< 548 »>. 




20 


t 386c. 


3. 


6 


e 563d. 


10. 


4 


ew 447 c. 


2. 


5 


e386a; ei415b. 


4. 


14 


ew 484b, 567d. 


11. 


8 


w 411d. 




6 


t 406b. 


6. 


33 


mt 351d. 


86. 


6 


t 388b. 


3. 


9 


m 428b. 


1. 


33 


to 524d. 




16 


to 524b,d, 532c. 


4. 


2 


to emended 506c, 


8. 


12 


mw 420d. 




30 


to* 348ab. 






348a. 


13. 


17 


e 312d. 




42 


e 422d. 


5. 


7 


m 535bc. 


19. 


2 


w 486d. 


88. 


2 


e« 348d. 


7. 


2 


ew348d;cf.408c. 




37 


t 403b. 


93. 


4 


w 506n. 


16. 


8 


ew 502c. 


25. 


L7 


ew 414d, 503d. 


96. 


2 


em 332d. 


18. 


3 


e 362c. 




18 


mw 335c. 




15 


ew 406a. 




7 


w 472b. 


26. 


6 


em 416d. 




21 


mi 408a. 


19. 


5 


et 357b, 356d. 


29. 


8 


et 331c. 




24 


e< 362a, 356d. 


25. 


1 


ew 457a. 


31. 


12 


to 526ab. 


91. 


1 


w 495d. 



Citations.] 



Noun- Inflection in the Veda. 



589 



RV. 


ix. 






RV. 


X. 






RT. 


X. 




97. 


16 


e 334d-5a 




47. 


6 


mt 552ab. 




94. 


12 


*401d. 




37 


ew 41 led. 




50. 


4 


mt 429b. 




95. 


2 


e 358b. 




38 


1 425a, 






5 


mt 357a. 






9 


e362d. 




53 


* 358a. 




51. 


2 


tow 408a,416c. 


12,13 


ew 505bc. 




54 


ew 392a ; 
534d. 


* 531a, 




4 
6 


tow 408a. 
TO 375d. 




99. 


7 
11 


ew 570ab. 

etw 488d. 




57 


w 483cd. 




53. 


5 


* 345bc. 




101. 


11 


mt 387cd. 


98. 


6 


e 378b. 




55. 


5 


e 336a. 




102. 


1 


« 386cd. 


101. 


12 


m 346a. 




57. 


3 


t 430c. 




104. 


10 


e 358a 


107. 


20 


e 335c. 




59. 


1 


ew 516b, 


520d, 


105. 


9 


mt 536a. 


108. 


11 


e 433b. 








534bc. 




106. 


3 


< 403c. 


109. 


7 


ew 506d, 


501e. 


61. 




t 481bc. 




107. 


3 


* 381d. 




14 


m 333d. 






9 


e 488b. 




108. 


3 


e 332d. 


112. 


4 


<341ab. 






12 


et 551b, 559b. 


109. 


1 


e 562b. 


113. 


10 


w 481c. 






21 


w 464cd. 




111. 


4 


e 466d. 










64. 


7 


TO<371d. 




114. 


10 


to 338b. 


RV. 


X. 








14 


e 354a. 




317. 




<341b. 


5. 


5 


tw 526d. 




65. 


10 


mt 552ab. 






9 


t 341ab. 


6. 


1 


mt 536a. 




66. 


9 


to 542ba 




119. 


13 


e« 330ab. 




5 


mt 427a. 




68. 


10 


w 485c. 




120. 


2 


t 373a, 377c. 




7 


to 524d. 




70. 


5 


et 415b. 




121. 


8 


to£ 354cd. 


10. 


1 


e 336c. 




73. 


6 


et 363b, 447b. 


123. 


6 


t 386d. 




12 


1 408a. 




75. 


3 


em 385b. 




125. 


8 


* 533b. 


12. 


2 


* 346d. 






6 


e 358c. 




128. 


7 


< 430c. 


14. 


7 


t 509b, 537d. 




9 


ot« 388a. 




131. 


7 


t 386d. 




11 


* 427cd, 565a. 


76. 


5 


e 534b. 




132. 


2 


TO* 341d. 




12 


i 482bc. 






6 


e* 562c. 




134. 


1 


tf 399a; ew 541d. 


18. 


14 


tw 403b, 455d. 


77. 


1 


w 521c. 




136. 




t 341ab. 


20. 


3 


to emended 506c. 




2 


to 345d. 




137. 




< 341 ab. 


22. 


5 


mt 341d. 






4 


efa) 521c. 






1 


w 512d. 




9 


t 373a. 






8 


to 524d. 




140. 


3 


e 384d, 395cd, 


23. 


5 


e 562b. 




78. 


2 


to 345d. 








363a. 




6 


TO 338b. 






3 


to 345d. 




141. 


4 


e 336d. 




7 


e 336b. 






6 


t 345c. 




142. 


1 


e 353b. 


24. 


1 


e 409a, 412a. 


79. 


7 


e451d. 






5 


to 338a 


25. 


4 


e 335b. 




82. 


3 


<444b. 




146. 


3 


e 376d. 


26. 


1 


TO 362b. 




83. 


6 


e 375c. 




148. 


2 


to 334a. 


27. 


20 


t 341ab. 






7 


< 342b. 




150. 


4 


ratf 388a. 




21 


e 534bc. 




84. 


1 


< 520b. 




152. 


2 


fa> 477ab. 


28. 


10,lU445a. 






v 


TO 376a. 




153. 


1 


et 362c. 


29. 


1 


ew 505bc. 




85. 




<341ab. 




160. 


1 


e 336d. 


30. 


14 


et 363ba 






5 


<493d; ea 


495b. 




3 


w 473a. 


31. 


9 


TO 556d, 


485d, 




24 


t 386cd, 557c. 


165. 


1 


< 385b. 






44tid. 






37 


< 557a. 






3 


< 389c. 


34. 


14 


* 386c. 






43 


t 376d. 






4 


t 386d. 


35. 


10 


mt 429b. 






46 


to< 404a, 


420c, 


178. 




* 341b. 


37. 


10 


w 485c. 








4264 






2 


t 341ab, 386d. 


39. 


1 


ew 546c. 




87. 


3 


<342b. 




183. 


2 


e412a,404a,413c. 




8 


tow 516a 






16 


e 359c. 




184. 


2 


* 544c. 


40. 


7 
8 


etw 529d. 
em 330d. 






17 
19 


ot< 535c. 
tw 370c. 




Val. 








9 


et 348c. 




90. 




< 47 lab. 




2. 


10 


ew 37 7d. 


42. 


7 


to* 548ab. 






10 


tow 509d. 




8. 


4 


e 353b. 


44. 


9 


« 386c. 






14 


fa) 370c. 




11. 


3 


mw 393c. 


45. 


12 


mt 548b. 




91. 


13 


w 473d. 






4 


TO 338b. 


46. 


5 


e 353c; TO 481a, 


92. 


9 


mw 540d. 






6 


i 397a. 






501c. 






14 


ew 523c, 


3a, 8a, 


AT. 


i. 






6 


to* 388a. 








551d. 




18. 


2 


w 474b 




7 


et 410d, 


414c ; 


93. 


1 


tow 391c. 






3 


< 412a. 






fa) 458d. 




94. 


11 


< 345b. 




20. 


2 


ew 418c. 



VOL. X. 



78 



590 



C. R. Lanman, [Index of Citations. 



AT. 


ii. 




AT. xi. 






TS. 




3. 


6 


w 482c. 


1. 2 


* 345b. 




viii. 5 


w 447ab. 


6. 


5 


to 477d. 


26 


e 337a. 




10 


w 552c. 


27. 


7 


w 491b. 


3. 19 


e 562b. 




xi. 29 


t 533b. 


35. 


4 


< 410a. 


4. 2 


ml 554a. 




xviii- 54 


ew 560d. 


AV. 


iii. 




AT. xiii 






xix. 2 


to516d, 512b. 


19. 


1 


t 406d. 


2. 33 


ew 363cd. 




xx. 2 


t 468b. 


20. 


6 


e« 336d. 


3. 16 


ew 415a. 




xxi. 33,57 


;'ew 560c. 


AT. 


iv. 




AT. xiv 






xxiii. 5 


ew 513c. 


3. 


3 


< 432d. 


2. 38 


t 557a. 




xx vii. 14 


w 466a. 


4. 


4 


mt 348c. 


AT. xvi 






xxx. 14 


w 408a. 


24. 


4 


cm) 362d. 


4. 6 


to 557b, 468b. 


xxxiii. 11 


e 433a. 


25. 


1 


mt 403d, 413b. 


AT. xviii. 




xxxiv. 26 


tw 516d, 512b, 


AT. 


v. 




1. 57 


to 521d. 






559c. 


1. 


3 


e 384d, 395cd. 


2. 12 


w 474a ; t 


565a. 


48 


e< 552cd. 


4. 


10 


mt 558b. 


3. 23 


t 362d. 




TS. i. 




19. 


2 


mj 506d. 


4. 60 


to* 524b, 532c. 


3. 2 


ew 481a. 


22. 


11 


* 495c. 


AT. xix 






14« 


e 433a. 


23. 


12 


mt 557ab. 


9. 4 


mw 542b. 




5. II 3 


t 564c. 


24. 


3 


* 430a. 


35. 5 


to 523c, 538d. 


8. 14 1 


< 468b. 


25. 


3 


<544c. 


37. 2 


? * 412a. 




TS. ii. 




27. 


6 


ew 420d. 


38. 1 


ew 372a. 




2. 6 s 


ew 506d. 


AT. 


vi. 




55. 3 


* 412a. 




3. 10 1 


w 466a. 


4. 


3 


« 432d. 


60. 1 


1 414a. 




TS. iii. 




23. 


3 


w 482e. 


61. 1 ! < 


J t 412a. 




2. 8 4 


w 447ab. 


35. 


2 


ew 567cd. 


AT. xx. 






3. II 6 


e 337a. 


41. 


1 


mt 383bc. 


136. 2 


« 341c. 




4. 4 1 


* 3f 7b. 


52. 


3 


w 470c. 








TS. iv. 




59. 


2 


* 377b. 


ST. i. 






1. 8 1 


w 466a. 


96. 


1 


mt 371d. 


175 


ew 362c. 




7. 13 2 


ew 560d. 


133. 


4 


m< 424b-6a. 


340 


e< 353c. 




TS. v. 




AT. 

48. 






557 


* 532c. 




6. 4 1 


to 653a. 


vii. 

1 


e 337a. 


ST. ii. 

477 


t 349ab. 




TS. vii. 
4. 9 


w 466a, 523b. 


81. 


1 


< 348a. 


502 


* 532c. 




5. 2 2 


w 497d. 


P7. 


1 


* 513b. 


1173 


* 398d. 








AT. 


viii 










Qat.Br. xii 




2. 


28 


t 406d. 


TS. 






5. 2 8 


ew 512d. 


3. 


17 


»ii 535c. 


ii. 20 w 569d, 


573a. 






AT. 


X. 




iii. 


5 t 533b. 




TBr. i. 




6. 


27 


t 381d. 


vi. 17 w 404bc 




1. io 2 - 5 


to 554ab. 



Sanskrit Index.] Noun-Inflection in the Veda. 



591 



SANSKKIT INDEX. 

The order in which the different classes of stems are treated in the several 
sections of the foregoing article may be seen from the Synopsis after the General 
Index. In every section the cases and numbers follow each other in regular 
sequence. It will therefore be easy to find at once those words about whose 
stem and ending there is no doubt. 

This Index gives the words whose analysis is uncertain, whose form or con- 
struction is ambiguous, whose accent is peculiar (and of these alone the accent is 
marked), or which are in any way noteworthy enough to call for special mention 
in the body of the article. In case of failure to find a compound by its firBt 
member, look for the last. 



dnh-as (vi.3.1), 563a. 

atikasu, 567cd. 

akrttaruk, 460d, 464c. 

akka, 360c. 

akravihasta, 568d. 

akrd, 438a, 442b. 

aksh, 558b, 392c. 

akshan, -aksha, 527ab. 

akshan, -i, 523ab, 530d. 

aksht, 391b, 37lb, 558b. 

agru. 401c. 

agruvai, 404a, 409d. 

aghdy&'ndm, 418c. 

aghoracakshus, 571a. 

aghos, 520a. 

angira, -as, -ds, -e, 550d, 
548c. 

angiro, A.p.m., 566c,342d- 
3a. 

acodate, 506a, 508a. 

ajoshds, contract N. p., 553a. 

ajma, -an, 526c. 

atafh",, 441b. 

jiandrdsas, 362b. 

ataptatanH, 40 2d. 

atasi'ndm, 398b. 

atiyafdm, contr't A.s.,553b. 

Stmand, 408c, 341d, Addi- 
tions. 

atra'is, 350a. 

atharva(voc.to-van), 554b. 

atharvi', 527d, 368bc. 

adatrayd', 358a. 

adabdhdyo, 569d, 573a. 

adbhis, -bhyas, 483bc, 567a. 

adrivas, 519d, Additions. 

advesha, -as, 548b, 554d. 

adhard'c, 454d. 

adhrigdvas, 414c ; cf . 432b. 

adhvagat, 502a. 

anak, 494d. 

anadut, -vh, 498b. 

anadvah. 498b; -vdn, 499c, 
559d-60a. 

anarvlm-am, 523c, 526a, 
527d'; cf. 531d, 532b. 

anarvdna-s, 526ab. 

tmarva-m, 526b. 



anarvd, f., 526b, 527d. 
anarvdnam, 330b, 526a. 
anavi, 411c. 
and', 337a. 
andga, -as, -ds, 551a. 
andgdstva, 556c. 
a-nimish, 494b. 
a-nimisha, 494b. 
anumati, 371b. 
anushthu, 401c. 
anushihuyd', 408d. 
an&'cas, -cos, 455d. 
an&ti, D.s., 382d. 
anehd, 551b, 559b. 
anyddrn, 490b. 
apadhd, 447 a. 
apavdn, 484b. 
apas, ap-ds, 501b. 
apas, 481bc. 
apasu, 484b, 567d. 
apt, 369b, 385b. 
aprajastd, -ya, -dstva, 556c. 
aprajds, 556b, cf. 439b. 
aprati, ■%, 394c. 
aprabhu, 403d. 
aprdyu, 40 id, 403b ; -ushe, 

571a. 
apsaras, -d, 551b. 
abibhyushas, N.p., 511d, 

513c. 
abhikkyd, 440b, 447 a. 
abhibhmn, 403d. 
abhimdtinam, 378b, 543b. 
abhimdtisMhas, 500c, 498d. 
abhishti-pd, 381a, 443d. 
abhihruU, L.s., 389c. 
abhlrunam, 374a, 404b. 
abhirvam, 403b, 407c. 
a/muni, 412c. 
amuyd', 408d. 
amrtdsas, 362b,d. 
amba, 360c. 
amW, 371c. 
dyd, ay&s, ayd', 357d, 358d. 
ayd' (i.87.4), 559b, 358d, 

495a. 
ayds, 493a, 5a, 546bc. 
aj^'ai -dni, 462d,5c. 



ayvmji, 462d. 
ararhgamdsas, 363cd. 
arant, 37 Id. 
aram/<2»$, 3 7 Id. 
ardii, 371b. 

anfoa=d™<2, 529d, 534bc. 
ari, 400b, 371 c. 
arishtabharman, f., 528b. 
arm^ 372d, 385b. 
arundhate, 372a. 
arc-(J, masc, 459c. 
arcis, 568cd. 
arnavdm, 353c. 
aryaman, 523b, 532b, 537c, 

538d. 
arvan, V.s., 520b. 
on><in£, -vara, 516d,23b,29d. 
arvdtt, -dn, -d, 517b. 
arvdcas, 458d. 
asMa, 360c. 

avadyabhiyd', 381d, 477a. 
avani, 37 lb. 
avantyas, 395b. 
avayds, 463b. 
avasrshtdsas, 362d. 
<M>frofe, 359b, 466d, 468a. 
aw»a, 554a. 
ouos, 344b. 
apa», ap«a, 527a. 
apipjiJ, 372c. 
apirshdnds, 526d. 
Ofvdsas, 362b,d. 
astoi, 342cd. 
osa», see asrt. 
asafcat, 467a. 
dsapcai, 467a. 
dsofcatam, 507c. 
4«opco«^, 467a. 
asiimid', 368a. 
asftoi, 404a, 408a. 
osr/, 461d. 

asrt, 463c, 466a, 523b,530d. 
asthan, -astha, 527b. 
asthan, 4, 523b, 530d. 
asffW, 447 a. 
asihAri. 394e. 
asmrtadhru, 401d. 
asmrti, 381a. 



592 



C. R. Lanman, 



[Sanskrit 



asya (asia), 338bc. 

aha, ahan, ahar, ahas, ahna, 

523a, 621a, 486b ; N.8.. 

630d. 
ahardivi, 488b. 
ahd, N.p., 539ab, 540a. 
ahighnyds, 384cd. 
ahrayas, 372c, cf. 37 5c. 
dkhu, 401c. 
djayi f, 388a,b. 
djaras, -sd, 555a. 
djdni, 371b. 

dtd, 442b,c, 438d, 441b. 
diais, 452b. 
ddhid', 477a 
d'pas, 481b, 483a. 
dprk, 459c. 
apes, N.s., 375c. 
<J6ftw, 403d. 
dyds, 358d. 
<%&, d'yas, 569cd. 
d'yu,. transition-stem, 569 

cd. 571d. 
d'yw, as G-.s.n., 572cd. 
dyurdad, 470c. 
dyushu, 403d, 419a. 
<fyos, 344c, 361b. 
druja-s, 462c. 
dvayds, 463b. 
<Jf<M, <2p<J, 550bc, 494a. 
dfirdd, -dyd, 447b. 
dpuj/d', 408d. 
dpushdndsas, 362bc. 
dshtrt,' S8$c. 
ds, dsan, dad, dsya, 492d, 

523a ; N.s., 530d. 
dsannishu, 536c. 
<feay<2', 494b, 523a. 
dhuti, 371c. 
irf, 466b, 493bc. 
idas, 466b. 
t'rf<2, 493bc. 
Iddbhis, 483d. 
indrajds, s.n., 445c, 560cd. 
indrdvdyu, 342d. 
indrafca, 340a. 
indrasvant, 570c. 
indrdpMian. 523b, 537c. 
indrdvatas, N.p.,516b,521a. 
indhanvabhis, i., 528a. 
%aretf, 504d, Corrections. 
fy<4»S, 510b ; cf. 521b. 
iradhyai, 563ab. 
ird, 493bc. 
inn, 542a. 
ish, 493bc. 
tsAii, transition-stem, 494b, 

552d. 
ishani, 381a. 
ishaye, 483d, 494c. 
ishidh, 47 5d. 



ishukrtd, 468a,d, 466c. 

isfaxJs, 403b, 455d, 410b 

foAvat, 403b, 409d. 

«rf, 466b. 

Mra, 490b. 

fewmi, 504d, Corrections. 

tfd, 490d. 

teW, 356b, 445a. 

ukthafds, -(ds, 493a. 

m/MM, 335d. 

ukthaw, 350a. 

ukshan-as, 524d. 

ukshan, -uksha, 527ab. 

ukshan, vkshdn, ukshn, 523 

c, 524a, 532ab, 538cd, 

540bc. 
ukhachit, 472a, 474cd. 
uccd', -ais, 337b, 455b,457c. 
«<2, udaka, udan, 523a ; 

N.s., 530d. 
udac, 455b. 
!«»M, 403d, 412c. 
upafruti, 381a. 
Mpas, 496c, 545c. 
updbhrti, 381a. 
ubhay'd', 358a. 
ubhayddat, 505a, 506d 

■ddtas, N.p., 509d; -dam, 

N.s., 506d. 
wrd, as L.S., 412c, 411cd. 
wm, p.n., 415c. 
uruvyac, 455a, 460b. 
uruvyancam, 460a. 
w*i, s.n., 406d; p.n., 415b 
uruct, 455a. 
wvdruds, 403b, 411a. 
urviyd', urvyd, 381c. 
wvi', L.S., 389c; d.n., 413d. 
urvy&ti, 381cd. 
up, upan, 523a. 
Ufadhak, 499a,c. 
upand, 441b, 559b ; -dm, 

551c. 
upane (i.51.11), 449a, 551e. 
«pan«(Ti.20.11), 524d, 533d. 
ush (stem ?), 545c. 
ushadbhis, 566d. 
ushar, 420a,c, 427a. 
usharbhut, 499a. 
ushas, ushd (trans.-stem), 

551c. 
ush&s, ushds, 545d-546b. 
ushas, as G.s.f., 563ab ; 

A.p.i, 563ab, 566c. 
ushds-as, G.s. or A.p., 523d, 

546c ; -dm, ib. 
ushdsdnaktd, 546b. 
usras. 426b, 429c; -n',427a. 
uti, 371b. 
utt, I.s., 380c ; D.s., 382d ; 

I.p., 396c. 



■ar, -as, 486b, 523a; 

N.s.", 530d. 
ASiar, as Ab.s., 488a; L.s., 

488b; A.p., 474d, 488cd. 
urnamradas, -a, 554d ; -ds, 

neuter, 560c. 
urminam, 374a,8ab, 542c, 

3ab. 
rkva, -an, 526b. 
rkvan, -vanl, 523b. 
rjisham, 543b. 
rju, p.n., 415c. 
rtanibhyas, 372c, 397b. 
rtayd', 358a. 
rM, L.s., 411cd. 
rtdvas, 519d, 536d. 
rdhadrayas, 382b,4d, 431c; 

ci. 401c, 4104 
rbhukshan, 523b,532b,538d. 
rbhva, -an, -as, 526bc, 545c. 
rbhvdfl «-, 529b. 
rshtvas, 519d. 
rhant, 504d. 
ekapatnis, 400c. 
ekamanta, 348d. 
ekdyur, 570a. 
eto-J, 42 6d. 
efefs, 363a,c. 
e?j<2, ercd', 332-3. 
e»<te, 363a,c. 
enos, 344b. 
ebhis, 483c, 349cd. 
ema, -<ra, 526a 
emusham, A.s., 511d, 512d. 
em&shas, 511d. 
evaydmarut 444a. 
evaydvas, 519d, 536d. 
evayds, 451d. 
asfto, 356a. 
ota, -o&a, 554d. 
o/as<2, G-.s., 563c. 
ojiyas, voc, 519d, 515b. 
o»«, 372cd. 
omand, 524cd. 
oshadht, 371d-2a. 
ktikud, 47 led. 
kakubh, kakuha, 471c. 
kanikrat, 505c, 507a, 517d. 
kanindm, 364c. 
kaniyasam, 514b. 
kaprth, 470a. 
kamadyH, 402e. 
karafyahe, 479a ; cf. 441a. 
karkari, 371c. 
kartari, 426d. 
•karma, -karman, 526d. 
kalydnyai, 369a, 383b. 
kavitvand, -tvd, 335d. 
ftd&utf, 47 led. 
Ms, fo2s<2, 494b ; fo2s-<w», 

494c. 



Index.] 



Noun- Inflection in the Veda. 



593 



Mm, 377bc. 

Myant, 504d, Corrections. 

kiydti, 508d. 

kiyedhds, 382c. 

ktdrnn, 4.90b. 

Mrine, 374a, 543cd. 

Mrtid, 379(1, 381d. 

kivant, 504d, Corrections. 

kuhiSim, 404a, 408a. 

kfJcaddfU, 402d, 407c. 

-fcW, 466c. 

ftr'fo, 405c. 

krtv&nas, 523c, 538d. 

JrsWipras, 448c. 

kratu, 405c. 

kratumatd, N.d.m., 516b, 

520d. 
krandas, 554a. 
kravis, 568d. 
*r<2, cpds., 438a. 
4«tea, 329b, 334c. 
krimindm, 399c, 4l7d. 
kshU, -an, -d, cpds., see 

kshd. 
kshdpas, 482 b. 
kshapd'bhis, 483bed, 493c. 
kshd, -&, -an, cpds., 438a,d, 

40a,b,la,2b,c. 
kshdm, 556d, 485d, 446d. 
kshdma, -an, 530c, 560a. 
kshds, 556d, 485d, 452b. 
kshitdyus, 569cd. 
kshipdbhis, 483bcd, 493c. 
kshu, 401c. 

kshumati, A.p.n.,516b,521b. 
fofte, 448a, 534a, 440a. 
kshoni, 372d-3a. 
kshm*?,, -d, 448b, 440b,2b, 

485cd. 
khd, -an, cpds., see khd. 
kharajru, 401c. 
khd, -&, -an, cpds., 438a,d, 

41a,2b. 
khddinam, 374a, 378ab, 543 

ab, 542c. 
khddinas, 374a, 393a. 
khdri, 372c. 
khidvas, 519d, 536d. 
jr3, cpds., see jd. 
gahgd, 440b. 
ganaprlbhis, 372c, 396b. 
garbhinayas, 372a. 
Saswtoi," 432a, 373d. 
y<2, g>3, cpds., 437d,8a,b,d, 

40b,la,2b,3b. 
gdm, 431d ; gdvarn, 432d. 
girayas, Ab.s., 393a ; cf. 

384d, 41 Od, 426c. 
giribhraj, 462b. 
yjA, as A.p., 488d. 
-S«, 401c. 



guhd, guhd', 358b, 500a. 
g&rtavacds, neuter, 560c. 
go and cpds., 431-2. 
godhd, 440ab, 444c, 5ab. 
gondm, 432a. 
gobhis, accent. 421c. 
govapushas, 571b. 
0aarf, 373b, 389c. 
gnd, 440a,b; -ds, N.S., 355 

b; p., 451b,2b. 
gndvas, 518b, 519d. 
gndspatis, 448d. 
gmantd, 509b,d. 
jmto, 448b, 440a,2b, 485d. 
glau, 434a. 

gvd, -d, cpds., 438a,d, 442b. 
ghand, 334d-5a. 
ghrnd, 335b. 
fffcfnS, 379c. 

'ghriavdnU, 521b; cf. 510b. 
ghrtasnd, 438c. 
ghrtasnu, 402a, 413c. 
ghrtdnnau, 342d. 
-^Ano, 479ab. 
ghrans, 530a, 480a, 495b. 
cakran, 505bc, 507a. 
cakriyd, 391cd. 
cakrusham, A.s.,511d,5 I2d. 
cafestes(i.92.9),57la; -msM 

(ii.39.5), 573a. 
cakshos, 410a, 569a ; cf , 

412d, 468c. 
caturnd'm, 397d. 
cafesAparf,-jfi(Jd,472bc,474b. 
candramds, -mds, 547cd, 

555c. 
camasdfl, I.S., 335b, 328a. 
camti,, I.s.,409a; L.s., 412a. 
caranyA, 402c. 
carathdm, 353b. 
cardyai, 554c. 
carishnu' , 402c. 
carmamnd, -&, 438c, 9a. 
cdkan, 505bc, 507a. 
cdyamdnd, 368a. 
cdrS, p.n., 415c. 
cikitushas, as N.p., 513d. 
cikitvas. 519d, 513b. 
a'fe', 371c. 
cfe-a, 368a. 
ciirdyus, 571a. 
c*oto?/as,384d,395cd,363a. 
^'S, cpds., seey<t 
jagat, 504d. 
jahghd, 440b. 
jYmm, 400b, 37lb. 
jan&'s, 571c. 
janydsas, 362b,d. 
jaras,jard, 550cd. 
jarimdnas, Ab.s., 534bc, 

523d. 



/a\ y«, cpds., 437d,8b,d,40b, 

la,2b,3b,6a. 
jdgrvadbhis, 514a; cf. 511c. 
jd'itijdtti,'; 409a. 
y<te», 401c. 
jdmdtar, 419d. 
joto*', 394c. 
jdyamdnam, N.s.1, 356c, 

348d. 
jdspati, 448c. 
jigatnU, 413b, 390d. 
jighaistf, 402c. 
,?ar, 416b. 

jeman-d, 523c, 537c. 
joguvdm, 373d, 404b. 
/«<*, cpds., 438b, 440b,c. 
,/«<&?, 492d. 

-j'fla, 401c. 

j'toos, 448b, 440a, 2b, 485d; 

-d, 447a. 
y^d, cpds., 440b,la,4d. 
jyd'ks, 432d; cf. 494d. 
jydyas, voc., 519d, 515b. 
,/^oi, jYoft, 456d. 
jyoMs, jyoti, 568d. 
jrayas, 554d. 

tofoa, -an, 526bc. 

takvavis, 373b, 393ab. 

takshan, 523c, 532a, 538c. 

tatanvat, 512c; cf. 511c. 

to«os, 479c. 

fcm<J, 479cd, 480bc, 515d, 
518d. 

tanu, 402b. 

tend, 404a, 413c, 412a. 

tanum, 404a, 408a. 

tanvd^m, 404a, 412a. 

fa»»$, 41 Id. 

tapani, 373b, 381d, 409a. 

tapas, I.s.n., 562b. 

tapushd (ii.34.9), 571ab. 

tapus, -os, 569a. 

tapojdn, 451d,2a. 

taranitvd. 335d. 

taras, 554d. 

tarubhis, 569b. 

fa«x2-g , am, 551d. 

tdvishi, 368b. 

tavishtvas, 519d. 

tasthushas, Corrections, 513 
a,c,d. 

tew, 419d, cf. 428c,d. 

Ugmatejasa-m, 555a. 

titirvas, 519d, 513b. 

iiropc, 455a ; ttrafcatd, 457 
c ; tirafcd', 455b,7c ; fo- 
rafcj, 455b,8a ; Tirafd', 
455b; 367d. 

tilapirijid', 381d, 369a. 

Usfnd'm, 397d. 

tvjaye, 462d. 



594 



C. B. Lanman, 



[Sanskrit 



turi, 37 lb. 

twigra, 438d,42b. 

tuvishmant, -tama, 570c. 

tuvishmas, B20b. 

tuvtravdn, 559bcd, Slid. 

trtsu, 570d. 

tend, 332-4. 

tofdsd, 545d. 

fonan, tmdn, 5230, 532ab. 

tmand, 408o. 

tmanid, f., 528b. 

tfr$, cpds., see fr<2. 

trasadasyavi, 411o. 

foyi, £ni, cpds., 437d,8b,41a, 

2b, 6a. 
trdmane, 524d. 
tridhdtU, p.n., 415c. 
trivishti, 381a. 
trac-i, masc, 459c. 
tvadrik, 456d. 
tvashtdr, N.S., 424b-5a. 
tvd-datta, -ddta, -hata, 334c; 

tveshita, tvota, ib. 
ds, cpds., see At 
dakshas, 551b. 
dakshi, 390a. 
dakshind, 358a. 
daksku, 570d. 
dadushas, as N.p., 513d. 
dadhan, -i, 523b, 530d. 
dadhanvatas, 51 6d, 513ab. 
dadhikrd, -krdvan, 523b. 
dadki'cd, 45 5d. 
Awffw*, 495b, 494d, 498a, 

499d. 
dara, L.s., 480& 
Ami, 505d; damt-am, 507b. 
dabhas, 485a. 
darwia, -cm, 526c. 
dafdnd'm, 397d. 
dasyavi, 411c. 
AJ, Ai, cpds., 437d,8b,d, 

41a,3b. 
Ata<2, 533b, 335ab, 330d. 
d&mame, -an, 524d. 
AJru, 401c, 408c. 
A?i«i»e, 524cd. 
AJfcrf, 505ab. 
<&2p<«, 490d. 
Afe, 492d. 
Afeai, 505ab. 
didrkshu, 405d. 
didyu, didyut, 466c; didyttt, 

368a. 
A'A/o«, 468b. 
Apsa[V]fo, 508c. 
dip, 432b, 489b. 
divakshds, 556b,c; cf. 439b. 
c?tv&, 466c. 
didivas, 519d, 513b. 
dlrghaprut, 474d,502c,503d. 



dtrghdyus, 569cd ; -<%o, 

572-3. 
div, stem, 489b. 
-<fora, stem, 486c. 
duras, as N.p.f., 486b. 
dwmitrdsas, 362bc. 
duhf&sus, 569b. 
dtisfitano, 41 2d. 
duhitar, 420a. 
<ZM, 370d. 
d&tidm, 389c, 313b. 
d&rebh&s, 495a. 
■drhk, -drhk, -drn, 490b. 
-A-p, -drp'o, 489d-90a. 
drfdye, 490a. 
<fc, 449b, 442ed,7e, 485e. 
devagopd, 439c,d, 445b. 
devatdtd, 386a, 466d. 
devayati'ndm, 399a. 
<foj)dr,419d-20; Additions. 
devavyacds, neuter, 560c. 
devahuti, 311b, 370d. 
devdn, G.p., 354a. 
As»&, N.s., 377b. 
doshasas, 557b, 468b. 
Aw, doshcm, 493b, 523a. 
dohas, 557d. 

dydm, dydvam, etc., 432c. 
A/«, 432b, 466c, 401d 

(cpds.). 
dyugat, 502a. 
Ay««, 466c. 

dyutadydmdnam, i, 528b. 
dyurnant&s, A.p., 521d, 

516b. 
dyumas, 520b. 
A/A, 489b. 

A/o and cpds., 432-3. 
A/cmW, 432d ; cf. 494d. 
A-a, cpds., 441a,4b. 
dradhas, -a, 554d. 
drau&t, 507b. 
dravinodas, -d, 556ab, 564c, 

449c. 
dravinodas-as, 563a. 
draviinud', 408d, 404a. 
drahydt, 507b. 
drdghmd, 533c. 
drdvinodasd-s, 563a. 
*•«, 401c, 408bc. 
-draft, 498a. 
druhyavi, 411c. 
dvdra, stem, 48 6e. 
dvdras, as A.p X, 486b. 
tfw'W, 358b. 
dviptid, -pad, 470d,2bc,3b, 

4b. 
dvibarhas, -a, 554d. 
dvibarhds, neuter, 560c. 
<ZA($, cpds., see dhd. 
dhakshat, 505a. 



dhakshi, 390a. 
dhakshu, 570d. 
dhanasdtayit, 388a,c. 
dhanasais, 452b. 
dhanutarau, 420d. 
dhanus, 569b. 
dhartar, 421d; -ro, 422d. 
dharma, -an, 526c, 527a. 
dM, tffia, cpds., 437d,8b, 

9a,40b,la,3b,4d,5a. 
-dh&ma, -dhdman, 526d. 
dhdsind, 381b. 
-AW, 367bc. 
<K#&, 403cd. 
-<fcj;, 499a. 
dhurtias. Ab., 383d. 
dhrshnuas, 410c. 
dhrshnuyd' , 40 8d. 
dhend, 358c. 
dhmdtar. 421d; -H, 423b, 

426d. 
Am/A 440b. 
-dhruk, 499a. 
dhruvas. 557c. 
naktabhis, 541a. 
naktayd', 358a. 
raaM. 371b. 
naktoshdsd, 546b. 
nodi, 370cd, 373b, 376bc. 
nadbhyas, 478a, 483c, 501e. 
nadyas, 394a, 373b. 
nandndar, 420a,c, 426d. 
napdt, naptdr. 419d. 
rcap£2, 373a. 
n&bJias, 484c. 
namasyd, 358c. 
mwji. 367d. 
mar, 420a, 421c, 428^ 
nar-as, G-.s., 426b,c ; N.p., 

as trochee, 428c. 
warA;?', 454d. 
rearAre, 373d, 430c. 
navagat, 502a. 
navdnd'm, 397d. 
navedds, contract N.p., 549d. 
navedds (i.79.1), 565d. 
»a»2/asi?Mjm, 515c. 
?»as, see »ds. 
nahus, -sha, 570ab. 
ndbhas, 484c. 
«dW, 371b. 
ndbhind, 381b. 
ndbhyds, 370c. 
ndras, 428c. 
ndrdfansi'ndm, 399a. 
raarf, -«, 370b, 371d, 399d. 
ndvdyd, 434a, 483d. 
ra<2s, »cte-A 494ab, 493a, 

546bc. 
»dsA -e, 361a, 494ab. 
nikdmabhis, 540d, 527b. 



iudex.] 



Noun- Inflection in the Veda. 



595 



nir&ak, ninik, 461d, 463c, 

456cd. 
ninios, 392c. 
nitikti, 381a. 
niddyds, 471bc. 
nidrd, 440b, 444d. 
nidhd, 440b. 
nimisha, 494b. 
niyayinam, 542c, 543b ; cf . 

378ab. 
niyutvas, 519d. 
nirrtt, 371b. 
nishkrti, 371b. 
»&:<?', -<?(, -<m's, 337b, 455c, 

457c. 
nrnam, 430. 
nrmands, G.S., 551d, 549c. 
»r», Gkp., 430cd. 
nemannish, 536c. 
«oa, 433d; -bhis, 421c. 
,p<5, cpda., see pd. 
padbhis, 475ab, 492a. 
paii, 400b. 
patnayas, 372a. 
pa<ft, pathi, 470a. 
pathishadi, 474a. 
patheshthd, 468c. 
.pad, 47 lab. 

^><Kfc[[m], as A.p.n., 349a. 
pados, as abl., 474b. 
panthanam, -as, 441cd, 532 

b, 538d. 
pcmtM, 441c-2b, 470a. 
panvd', 468d. 
pararneshthina-m, 542b. 
pardcais, 457c. 
pa/rijman, 522d, 524a. 
parijmdnas, i, 528c. 
^>arir<Spas, -rdpas, 482d. 
parihvrtd', 502d, 477a. 
parasto?,368b; -iste, 372b. 
paMtid', 369a, 381d. 
pdliknt, 368b. 
pavayd', 358a. 
paviravi, 411c. 
j»apa, 405c. 

papumdnti, 52] b; cf. 510b 
papush&s, 444ab,8c,51d. 
papcd', -<M, 337b, 455b,457c. 
;papjja-isiW$, 380d. 
papud', du., 403c. 
pashthavdt, 499b. 
pastios, 344b. 
pasprdhdndsas, 362bc. 
pd\.2>a, cpds., 438bc,9a,40b, 

la,4d. 
pdih-as, 470ab. 
pdda, 47 lab. 
pdrthivdsas, 362bc. 
pdvakafoce, 568c, 572d, 

390a. 



pdvakdsas, 362bc. 

pdshios, 344b. 

p«'tor and cpds., 419d-20. 

£>«'fa, 405c. 

pitr-os, 420c, 428b. 

piiras, G.s., 426bc. 

pipru, 401c; -mos, 410c. 

^)«»s, 493a. 

puTiarbhuvds, 404a, 409d. 

411a. 
pumas, 519d, 496c. 
pumdns, 493a, 495a. 
pwa, stem, 486c. 
puraetdr, N.s., 424b-5a. 
pwramdM, 371b,d. 
pwras, 488d. 
pwS, p.n., 415c. 
p«njdamp<j, 559b. 
purudansas, -a, -d (contract 

du.), 551d. 
pwnmishthds, 556b, 439b, 

43b. 
purupefdsu, 548c, 554b. 
purtibhu, 403d. 
puruh&td, 342d. 
pwu, s.n., 406d; p.n.,4I5b. 
p«r4c?, 456a. 

puroddp, -ddfa, -das, 489d. 
pwrodds, 490c, 463c. 
purodhas, 556a. 
pushti, 371b. 
p&rvidm, 357b. 
pilshan, 523b, 532b, 537c. 
p&shan, pUshana, 526d. 
jritaAd a-, 529a. 
prtandshdham, 499a. 
prtsushu, 469d. 
prthivi, du., 391c. 
pr**, 37ld, 367d. 
prihwjraya, 554d. 
prthuydman, f., 528b. 
prthushtu, 40 Id. 
prddkuds, 404a, 411a. 
^rpa^, 372b, 394b. 
PfP™*, 371b. 
prfnigdvas, 432b. 
2>j-sft«^, 391b, 509c. 
prshant, 504dr 
prshtha, 440d,la. 
2>rsft*M, 37 lc. 
pepos, -ds, 554b, 555b. 
^epos, I.s.n., 562b. 
pawrU, dual, 342c. 
prS, cpds., see .pre!. 
prdcetasdndm, 555a. 
pracetd, as voc, 564c. 
-prajds, transition-stem,556 

be. 
prajd, 440b. 

2>ran«$, 380c, 383a, 389a,c. 
pratidhd, 447a. 



praUdhdm, 445c,7d. 
pratishthd, 447b, 363b. 
pratisadrh, 490b. 
pratt'eas, 4m, -cas, 455d. 
pratyahk, 456a, cf. 490b ; 

-aSci, 459a. 
prathind, 533b. 
prabhu, 403d. 
prame, 447cd. 
prayatsu, 466a. 
prayasvantas, A.p., 52 led, 

516b. 
prayukti, 381a. 
pravadydmand, 524cd. 
prahd, 440c. 
pra-hva-s, 486b. 
prd,pr&, cpds., 438c,9a,41a. 
prdnk, Additions to 456a ; 

cf. 490b. 
prdcd',-08, -aw,455c,457c,d. 
prdtaritvas, 519d, 536d. 
pnJ'f*, 490a, 491b. 
^w<%jft, 402d. 
preniJ, 533b. 
preshd', 494b. 
proshiisham, A.s., 511d, 

512d. 
psoras, 551b. 
-psw, 401c. 
bandhv/r, 486d. 
JaM, p.n., 415c. 
bahuprajds, 556b,439b,43b. 
bahvind'm, 398d. 
btbhatsu, 402c, 403d. 
brhadraye, 382b, 431c, 384 

d ; cf. 401c, 410d. 
brhant, 504d. 
brahrni, I., 528b. 
bJiaktivdnsas, 513c. 
bhagavas, -gos, 520a. 
bhanjatind'm, 398d. 
6AooVopoce,568c,572d,390a. 
bJiaras, 557b. 
bhavas, 509a, 559d. 
Ws<2, cpds., 440bc,4c,6d. 
-6*4;, 462b. 
bhdttnd'm, 398d. 
bhdnumas, 520b. 
SMs, bhdsvant, 492d. 
bhiyai, 373a. 
bhishd', 561d. 
6Ms, 492d. 
Sfttflfi, 371b. 
bhvjmti, 530a. 
bhutijatind'm, 398d. 
-6M, cpds., 401d, 403d, 

405d, ff. 
6M«4, 533ab. 
bh&mand, -as, 524d. 
JMrad (vi.62.8), 539d ; cf. 

388c, Additions. 



596 



G. R. Laninan, 



[Sanskrit 



bhOmi, 371b. 
bhAmids, 310c. 
bhuri, 394c. 
bhrti, 371b. 
hhejdndsas, 362bc. 
bhos, 509a, 559d, 520a. 
bhraj, bhrdj, 462b. 464b,c. 
bJirdtarsmd cpd., 419d-20a. 
makshU, 402d. 
maksh&bTiis, 41 6d. 
maghavan, -vant, 516c, 523 

b, 529d. 
maghonas, as N.p., 523c, 

538d. 
madrik, madriadrik, 45 6d. 
madhu, 405c; p.n., 415c. 
madhti', madhvt, 402bc. 
madh&yuvd, 403bc. 
madhyd', 358a. 
madhydywas, 403bc. 
manas, I.s.n., 562b. 
mandyai, 359d,a. 
mantshd, 356b, 445a. 
maim, -its, -usha, 570d,b. 
■manushydn, G-.p., 353d. 
manotdrd, 420d. 
rnantwmas, 520b. 
manthd, 441c-2b. 
mandim, 542c, 543b. 
mayobhU, 403d ; p.n., 415c. 
marutvas, 519d. 
maridn, G.p., 353d-4a. 
martdsas, 362b,d. 
marydh, 339d. 
maryddd, 445b,52a. 
mate, 498d. 
mahant, 504d. 
mah&ye, 498d. 
m&has, mah-as, 501b. 
mahas, I.s.n., 562bc. 
mahdtmdnas, A.p., 523d, 

540c. 
mahdn, -dn", -aan, 506c. 
mahdnasa-s, 555a. 
rooMwfc', 510b. 
mah-dm, as G.p., 501c. 
mahdm, as N.s.m., 506d; 

as A.S., 552a, 442a, 532b. 
maft?, du., 391c, 392a; p.n.. 

394d. 
mahitvand, 336ab. 
mahitvd, 335c-6a. 
mahind, 533ab. 
mahimanas, 524d. 
mah-i', 500d, 499a. 
mafci(-«2),356cd; L.s.,389e. 
roaMuJ'ro, -?'n<2m, 398d- 

9a, 541d ; cf. 533ab. 
mahiyai, 359a, 447d. 
aA«4, 533b. 
m <2, cpds., 440c, 444d,5b. 



mdhs, -so, 494c. 558bc. 
mdtar and cpds., 419d-20. 
mdtaribhvari, -pvari, 559d. 
mdtaripvas, 519d, 536d, 

559d. 
mdtalt, 367d. 
mdMr, N.s., 424b,d. 
mdtr-os, 420c, 428b. 
-motf, 47 Od. 
mddbhis, -bhyas, 567a,493b, 

497c. 
mddhvt, 390d. 
m<%ft, 412a. 
ra<2s. 492d, 493b. 
mds, mdsa, 493d, 547cd, 

496bc. 
mdsu, mdssu. 567d, 49 7 d. 
mitadru, 401d. 
mitrakrH, 402a,d. 
mitrardjdnd, 342d. 
mitrdyuvas, 403bc. 
mithuyd', 4084 
mith-a, s.n., 406d. 
mtdhvas, 519d, 513b. 
medhas, -d, 552abc ; cf. 

549c. 
mefta#«M<2, 404a. 
yakan, yakrt, 523b, 530d. 
-j/o/, 462b.' 
yajnamush, 493a. 
yajnavacas, I.s.n., 562b. 
yajHd, 334c. 
yatind'm, 398d ; -fo4»i, 

399a. 
^aratas, 510b. 
yantwr, 486d. 
j/a^i, 371c. 
2/aArf, 372b, 396a. 
V<2, cpds., 438c,40c. 
yoVS, 355d. 
ydt, 464cd, 462b. 
ydma, ydman, 526d. 
ydmanas, 524d. 
-i/M, 401c. 

2/»im, 456b, 462a,3b. 
j/m«/, 462a,3d,4d,5c. 
yuvat, 530d, 560a ; -atas, 

516cd, 540b. 
yuvati, 371b, 372a. 
yuvadrik, 45 6d. 
yuvan, -vant, 516cd, 530d, 

522d. 
yuvayHs, 415b. 
3/w<Jfctt, 406a, 415c. 
y&thidm, 353b. 
?/fc<2, as du., 523c, 537d. 
y&'ne, -as, 408c, 527c. 
yi!a, y&sha, -an, 523a, 527a; 

N.8., 530d. 
yend, 332-4. 
yonayit, 388a,b. 



yoshdn-as, (<$) 523b ; fern. 

528b. 
yoshan, -d, -and, 523b. 
^<w (for ydyos), 344b. 
rakshohanas, 478c, 481a. 
raghatas, 466a. 
raghudrli, 402a, 407a. 
raghuyd', 408d. 
rajishthais, 350b. 
rajjimi, 411c. 
rajjvdm, 404a, 412a. 
ratnadheyd, 336c. 
rathayw, 415b. 
rathirdyatdm, 506a. 
raftfe, 373bc. 
rathyd, 336b. 
r<™, L.s., 480d. 
rappad&dhabhis, f., 528c. 
royi, 431. 
rayivas, 519d. 
-ravas, 557d. 
rapmd, 533c. 
rasa, 360b. 
r<J, 440b. 
-r4/, 462b. 
rdjan, -raja, 527b. 
r<j*i -5, 370c, 372a. 
rdtJuy'iteyt'ndm, 399a. 
r<Jm. rdj/d, etc., 431 ; cf. 

432d. 
rdshtri, 367d. 
ripuas, 410c. 
rufant, 504d. 
retodhas, -d, 552c, 555d. 
revatayas, 372a. 
rote, 553c. 
reaVis?, rocfas, 558b. 
rrffa'pi, 368b; -Jste, 372b. 
ra», 431. 

lakshmi, 368d, 376d. 
vaktart, 42 6d. 
natoa, -a», 526bc. 
vaksht, 372c. 
vacas, I.s.n., 562ab. 
vajrivas, 519d. 
vadh&m, 404a, 408a; -aaf, 

409d. 
vadhtmantas, A.p., 52 Id, 

516b. 
jjarafi, as N.p.n., 348d. 
j)o»<Jm, 353c. 
vonjniini, 542bc. 
tiaraisfttftA 411d. 
vantis, 569b. 
vanejds, 45 Id. 
vandhur, 486d ; cf. 494c, 

558bc. 
«a»«4na»as, 362bc. 
vapushdya, 570c. 
wi^flw, 'fowls, fowl,' 557d- 

8a, 395b, 384d. 



Index.] 



Noun-Inflection in the Veda., 



597 



vayds, 'strengthening, '-d'm 
(contract A.s.), 552cd. 

vayd'h (vii.40.5), 552c,362d. 

vayodhas, -d, 555d, 564c, 
449c. 

vayodhai, 447d, 452bc. 

varind, 533b. 

variman, -mant, 516d,532c. 

vartitrayas, 372a. 

varcodhas, -d, 556a. 

variants, 377b, 395c. 

varpanitis, 554c. 

vaiarhruja-s, 462c. 

valagahanas, 478c, 481a. 

vavrdhantas, 505b, Correc- 
tions. 

vafa, A.s.m., 349b. 

vafdsas, 362bc,d. 

vashatkrti, 381a. 

vasu, 405c ; s.n., 406d ; 
p.n., 415c. 

vasudhdtSras, 420d. 

vastk, p.n., 415b. 

-vdh, 498ab. 

vdhat, 466b ; cf. 506a. 

vd, cpds., 440c. 

vdghat, 505ab, 506a. 

vdcas, 461b. 

vdjabharmabhis, f., 528b. 

vdnias, 393c. 

vdta-svanas-as, 546c, 345c, 
557b. 

vdyavas, Gr.s. ?, 410d; 414c. 

j>dr, «(Jn, 486c. 

vdprdsas, 362bc. 

-«>dft, 498b. 

^'(trndto - , 419d. 

vit, 490c; vidbhis, -bhyas, 
492a. 

viddndsas, 362bc. 

vidushas, N.p., 61 Id. 

viifos, 569b. 

vidyot, 468b. 

vidhartar, 42 Id; -«', 422c- 
3a. 

vidhartdr, U.S., 424b,d. 

vidharman, V.s.n. ?, 536d. 

vidhavdm, 330d, 543b. 

wpd?, 490c. 

vibhdva, -an, 526bc. 

vibhdvas, 519d, 536d. 

otM«, 405c, 402bc, 414b, 
416d. 

vibhvan, 522d, 524a. 

s»'&A«ki»(J, as Ab. 1, 534b. 

vibhvan" r-, 529ab. 

vimrdhas, 477ab. 

vUiptids, 369a, 385b. 

vic&', 402b. 

vipvakarma, -an, 526c. 

vipvakrshti, 371b. 

VOL. X. 



vipva[m], as A.p.n., 348d. 
vipvavyacds, neuter, 560c. 
vipvd'rt, 454d. 
vifvdyu, 569cd. 
vipvdyuposhas-am, 548b, 

557b. 
vipvdshdt, 499a. 
vifvd?, 416c. 
vishundk, 456c. 
vishtapam, -d, 481cd. 
vishnavi, 411c; -n«ow, 410c. 
vishpat, 490c. 
vishvdc, 454d. 
fM«, p.n., 415c. 
viduhards, neuter, 560c. 
etoto, p.n., 415b. 
i)«2, D.s., 383a. 
rfrS, dual, 342c. 
vi/rapepds, neuter, 560c. 
viravas, 519d. 
«W(J, -e»?<2, -end, 331d, 336b. 
vrt, 466c. 
vrtayd, 483d. 
vrtraghne, 479ab. 
vrtra/iam, 478d, 480a. 
wfr-aMs, 442c, 47 8d. 
vrthSk, 456c. 
vrddhdyum, 569cd. 
vrdhas, 557d. 
vrdhasmt, 402a. 
vrdhdsas, 362bc. 
»r*M,asneut.,how avoided, 

'530d, 531a. 
vrsha, -an, 527a. 
vrshan-as, 524d. 
vrshatvand, -tvd, 336b. 
OTsft#», vrshdn, vrshn, 523c, 

' 4a,8c, 532a,b,'7c,d, 8c,d, 
540bc. 
vr'shand'ma, 531a, 534d. 
vrshabhd, 339d. 
wsM, fern., 528c. 
wrsAi?, D.S., 383a. 
vr'shni, -iam, 531a. 
i>«K, 389a,c. 
vedhas, vedhdm, 549bc. 
rapas, 557ab. 
veshand, I.s.n., 349b. 
veshaprts, 372c, 375b. 
fles, N.s., 375c. 
vehat, 466b ; cf. 506a. 
vyd^ghra, 432d; cf. 494d. 
vydhrti, 371c. 
vyitshi, 494d. 
vyomanas, 524d. 
w<&, 440a,b,2b,4c. 
patefe, 376d. 
pafozre, (M*r£, 523b, 530d. 
pafcifeow, 5194 
pakvarayas, 372a,d; 377b. 
fact, stem, 371c. 

79 



pactvas, 519d. 

paiakratu, 405c. 

patoitfW, 348d. 

cataydmnd, 524b, 532c. 

patard, 431c. 

patasds, s.n., 445c, 560cd. 

catdvan, 520b, 509a. 

panakais, 457d. 

panais, 457d, 585 (Correc- 
tions). 

famS, 4, 1.s., 381c; -?, p.n., 
394c. 

cambhu, 403d. 

cardha, -as, 553cd. 

pawis, I.s.n., 562b ; G.s.n., 
563d. 

cavasd, Gr.s., 563d. 

fa«as<toa», 520b. 

cacvatd, 518d, 480bc, 515d. 

cd'kin, 542a ; -inos, 542c. 

pdsai, 505ab. 

pdsd, 363b, 495d. 

p&?<2»» (ii.23.12), 353cd, 
495d. 

pikva, -an, -as, 526bc, 545c. 

fipu, 405c. 

firsha, -an, 526d, 527a; 
N.s., 530d. 

cirshd, N.p., 539ab, 540a. 

pukrapoce, 568c, 572d, 390a. 

pukrasadmandm, f., 528b. 

pucayas, 384d, 395cd. 

pucijanmanas, f., 528b. 

pjec&fan, 506c. 

p«a&Ar<J?(te,565d,585, Addi- 
tions to 345, 362. 

pucishmas, 520b. 

fimd, -e, -as, 527c, 408c. 

purudh, 475d. 

pushkasrotdm, contract A. s., 
553b. 

p&ghandsas, 362bc. 

pfljapdninam, 378b. 

prtapdre, 45 Id, 2a. 

pepas, 557a. 

perns, 557c. 

pokas, 557c. 

pom, 568c, 572d-3a. 

prat, prath, prad, 470a,466d. 

praddhd, 436d, 444d, 440b. 

prad-dhe, 447cd. 

psrarnayuvas, 403bc. 

prutarvd, 529d. 

pruti, 371b. 

prushti, -ih, 380d. 

prenidan, 506c. 

preshthasthds, s.n., 445c, 
560cd. 

(TOni, 371c. 

pwan, 522d. 

pvaninam, 374a, 378c. 



598 



C. R. Lanman, 



[Sanskrit 



poanibhyas, 372c, 391b. 
fvapruai, -ds, -dm, 404a, 

409d, 411a, 412a. 
fvasivdn, 512b. 
pviticayas, 458d. 
pvetavds, 463c. 
pvetid, 358c. 
shdt, 499b. 
s&, san, cpds., see siJ. 
sarhvavrtvat, 512c; cf. 511c. 
samviddndsas, 363bc. 
samsrdbhis, 483bc. 
samhat, 442c, 502a. 
sakthan, -i, 523b, 530d. 
sakthi, trans.-stem, 37lc. 
sakshanl, 390d. 
sakMy, -i, and cpds., 400 

be. 
sdkhyd, 336bc. 
samgati, 3 lib. 
sacathyaw, 350a. 
sajoshas, -ds, contract NVp., 

552d-3a. 
satyatdtd, 466d. 
satyarntigra,, 348d. 
saty&m, 357b. 
satrdshdt, 499a. 
satrdsdhe, 500a, 498d. 
satrdham, 47 8d, 480a. 
sdtvart, satvand, 526ab. 
sddaspaiA, 563b. 
sod-d, 470d. 
sadrnk, -drnn, -drn, 490b. 
sadro8 (i.173.3), 539d. 
sadra, 401d. 
sadhamdda, 471c. 
eadhamds, 472b, 463c. 
sadhastuti, 381a. 
sadhrias, Ab., 384b. 
sadhri, 380c. 
sand, -d'«, 337a. 
sap^Kd'm, 39 7 d. 
saprafhas, -a, 554ab. 
saprathds, neuter, 560cd. 
sapswrdsas, 546c, 551b. 
sabardhvm, 403cd, 407d. 
samaras', 358a. 
samit-samit, 476b, cd. 
samfcas, 455d. 
samudrt, 369b; -ij/as, 394a, 

396a. 
sammitdsas, 362b,d. 
samyaKci, 459a. 
samyat, 456d, 466a. 
sayugvd, I.. 527d. 
aayoni, 371b. 
»ara<, -adbhias, 454a,b. 
sarasi 389c, 558bc. 
sarasvas, 519d. 
sarah, 454b. 
sarin, 542a. 



sarvatanH, 40 2d. 
sarvatdtd, 466d, 386a. 
sarvwraihd, 336d. 
sarvdhrdd', 473 a, 477 a, 

358a. 
swrvdyus, 569cd. 
salakshmd, t, 528b. 
sdfcai, 467a. 

sapcaf, 467a, 466b; cf. 506a. 
sasthdvdnd, 1, 528a. 
-sSA, 498a-d. 
saftos, 55 7d. 
sahasracaksho, 569a, 572d, 

412d; cf. 390a. 
sahasrdddvndm, t, 528a. 
sahasraparnid' , 381d,369b. 
sahasrafodhu, 405c. 
sahasrarnMi, 348d. 
sahasvas, 519d. 
sahd'van, -vant, 516c, 523b, 

529d 
s<J, sfl, san, cpds., 438c,9a, 

40c, 42b,c. 
sddharanyd', 368a. 
sddhuyd', 408d. 
sdnavi, 387a, 412b. 
sdra«, sraa, 401c, 408c. 
s<?»S, p.n., 415c. 
sdrao, etc., 387a, 41 2b. 
sdrafe', 510b ; cf. 521b. 
sdrasa, sdras, 558b. 
-sdft, 498b-d. 
simd, 339d. 
sirf, 367d. 

sukarmahhis, f., 528b. 
sucetimam, -d, 404c, 407c, 

408c. 
sujanmomi, f., 528b. 
SMtor-mdnam, f., 528b. 
sutavds, 488d. 
strfrdmdraam, f., 528b. 
suddtu, p.n., 415c. 
s»dds, 492d. 
sudruam, 403b, 407c. 
supath, 470a. 
sapaptorai, 380c. 
suparni, 372b, 394b, 396a. 
suparad, 403d. 
suprajastva, -dstva, 556c. 
swprajds, 556b, 439b, 43b. 
SMpraraiiJ, 382a. 
suprayds, s.n., 445c, 560cd. 
soprdcj, 454d. 
saprdsid, 382c. 
sumedhas, -dm, 552ab. 
sumna, 441a, 2b. 
swdbhi, 394c. 
surabhishtamci, 570c. 
mrddhds, contract N.p.,553 

ab. 
suva/rcasa-m, 555a. 



suvdstvds, 403b, 411a. 

SMflrMi, 381a. 

swparaS, -?, 381c. 

swparmdnani, f., 528b. 

SMfasft', 381a. 

sush&md, t, 528b. 

smsWm, 401d. 

sms/i?Am, 401c. 

svhantu, 409a, 413b, 415c. 

svhavd, 336d-7a; -dA",335c. 

sSraawi, 411c. 

sm"V-os, -e, -e, 408c, 48 7d, 

8ab. 
sHrmt, -it, 369a. 
srrai, 371c; -id, 381d. 
sr'dhas, 47 7d. 
sendnibhyas, 372c, 397b. 
so, N.s.f., 356a. 
sobhari, 37 Id, 367d. 
somapa, -d, 436cd. 
somdptehan, 523b, 537cd. 
staryam, 394a, 379b. 
sfrnas, 421a, 428d ; -fctas, 

419d, 421c. 
striyas, N.p., 394b. 
sfe% 355c, 372b; -bhis, 

396d; -nd'ra, 398d. 
siftd, -a, cpds., 437d,8c,9a, 

40c,la,3b,6a. 
sthdtar, -rdm, -ur, -r, -re, 

421d-3c. 
sthdtdrd, as I.s. ?, 534bc. 
sthds, s.n., 445c, 560cd. 
snihiti, 372b. 
srn, 401c, 408bc, 416d. 
snehiti, 372b. 
spat, 490c. 

smayamdndsas, 362bc. 
syHmana, 524d. 
srawai, 466b; cf. 506a. 
sridhds, 47 7d. 
srucas, 461b. 
svaneds, 558a. 
svatavadbhyas, 566d. 
svatavdn, 559bcd 
svatavobhyas, 567a. 
swadftd, 440b, 444d,6d,7& 
svadhdvas, 519d. 
svanas, 557b. 
scarai, 377d. 
s»apa<2/ai, 359a, 447d. 
s»a? raa^d,T.s.m.,334b,358a. 
wdm, sud'm, 404a, 4] 2ab. 
suar, 487b ; as G.S., 488a ; 

asL.s., 488b. 
svaras, 557c. 
st>or», 369b. 
svarcis, 568d. 
SDarraar, 420a. 
svarshdtayit, 388a,c. 
svavdn, 569bcd. 



Index.] 



Noun- Inflection in the Veda. 



599 



svafods, 511a. 

svasar and cpd., 420d,& 

svasdr, N.s., 424b-5a. 

svasr-os, 420c, 428b. 

svasii, 381a, 383a. 

svasrdm, 430d, 423b, 373d. 

svdhdkrti, 371b,c. 

svend, 332d. 

M, han, cpds., 439a,41a,2c. 

-toi, 478-9. 

hantibhydm, 403c, 413d ; 

cf. 371c. 
Jianvds, 401a. 
A«n»j, 392a. 
haridru, 401d. 



harivas, 519d. 
Aanpmop<?r«, 534bc, 529d. 
haripriyam, 377d. 
havamdndsas, 362bc. 
havishkrti, 381a. 
havishmanlas, A. p., 521c, 

516b. 
havishmas, 520b. 
hastayos, as abl., 474b. 
hastini, -ini', 368a. 
hastinid's, 369a. 
Mriyojand, 339d. 
hinsdndm, 353b. 
Aimd, 485c, Additions, 585. 
hiranyakepiai, 369a,b. 



Mranyayd', 357c. 
Mranydyd, 381c. 
hiranyavdfts, 400c. 
Arrf, hrdaya, 471c. 
Arcfe", 473 d. 
ftr*'<S, 363b, 447b. 
hedas, 553c. 
Aeft', 371b; -y&s, 370c. 
fte& -sA, 380d. 
hemand, 524d. 
farter, N.s., 424bcd. 
homa, -an. 526c. 
hradecakshus, 57la. 
i, 373 c. 



GENERAL INDEX. 

This Index is intended merely to facilitate reference to the phenomena of a 
few classes, of which mention has been made here and there. Page-numbers of 
the same decade in succession have been abbreviated by omission of the first 
two digits. 



&, roots in, 434-7. 

■dft,, for -dns, 346bc; cf. 395a, 4]5d, 
429a, 499c, 506c, 512ab, 514bc, 5l7ab, 
559bc. 

Accent, differentiates homonyms, 358 
a, 368a, 414b, 454d, 455bc, 481c, 501b; 
modernized, 455d ; of proper names, 
368a, 454d, 455b; of transition-forms, 
483d, 434a, 462d, 490a, 493c, 494c, 498d; 
shifted in forming feminine adjectives, 
368-9, 402 ; shifted in case-forms used 
as adverbs, 337ab, 339d, 358a, 405d, 
408d, 433d, 455bc, 457cd, 458a, 473a, 
479d, 482b, 494c, 495c, 500a, 507b, 523a, 
555a, 561b,d. 

Adverbial shift of accent, see Accent ; 
case-forms used as adverbs, see Case- 
forms. 

as and a, relation of equivalent stems 
in, 546-58. 

Aspiration, transfer of, 47 5d, 499a. 

Atharvaveda, material taken from it 
for foregoing article, 325d, 327a. 

Augment, loss of, 356c. 

bbh, occurrence of, 483b, 585, Addi- 
tions to 485. 

p, inconsistent treatment of final, 490 
ab. 

Case-ending: dropped. 535-6, 387a, b, 
389c, 476d, 480d. 488b, 562abc, 563ab,d, 
566c, 572cd; doubled, 469d. 

Case-forms used as adverbs, 331b, 
333ab, 337ab,d, 358a, 364a, 377c, etc. ; 
their accent, see Accent. 

Catalectic pddas, which Grassmann, 



perhaps unnecessarily, makes acatalectic 
by resolution in the cadence, are men- 
tioned, 525ab, 338a, further, 336a, 345 
d-6a, 350ab,d, 352c, 357a. S61a,2b,3b,4c, 
371d, 381c,3d,4b,8a, 397b,d, 398, 399, 
417c, 418, 429d, 430ab,lb, 441d,3c,6a,9d, 
450d,2b,d, 469cd, 478a, 483d,4ab, 497c, 
504b,9c. 515c, 51 7d, 522a,9d, 535cd.8ab, 
541b,c, 559a, 567a. 

Confusion, of strong and weak forms, 
see Strong forms for weak, also Weak 
forms for strong; of dual case-forms, 
see Dual case-forms ; of genders, see 
Genders interchanged. 

Contract forms, 549-53 ; 545d, 555d, 
561bc,2a.3d,4a,5a,d,6c,7b; cf. gdm, rdm, 
dydm, gos, maghonas, and 432cd. 

Crasis after elision, see Elision. 

d: for bh, 47ld ; for s, 493b. 

d: for d, 475ab ; for f, 492a; for sh, 
493ab. 

Differentiation, of homonyms by ac- 
cent, see Accent ; of forms by transi- 
tion, 436d. 

Dual case-forms, confusion of, 343cd, 
344b, 361b, 392cMs, 474b. 

Ecthlipsis of to, 522c, 533ab; of n, 
533c. 

Elision and crasis, 329cd, 330ab,la,c, 
348bc,9b, 357b, 362a,3c, 375d, (cf. 376a,) 
378c,9c, 381a,4b,5b,7cd, 405d,6a, 425a, 
444a, 6d, 517d, 563cd. 

Equivalent forms, their abundance, 
405c. 



600 



C. R. Lanman, 



[General Index. 



Feminine formation, 367cd, 521061; 
with change of accent, 368-9, 402 ; 
double or triple, 402bc. 

Formal parallelism. 554ab,c, 468b, 
419c, 486d, 516b, 540d,3b,8abc, 568d. 

Genders interchanged, 516a : m. for 
f., 427d, 510b, 527d?, 528c?, 562bc; m. 
for n., 330c, Slid, 510b, 531a; f. for 
m., 343bc, 391b, 416bc, 515c, 516a; f. 
for n., 344c ; n. for m., 391b ? ; d. for f., 
344c, (cf. 391b,) 515d,6a,8d ; masculine 
declined with neuter ending, 433a; i 
with m., 416b, 429bc ; n. with m., 343bc. 

Genesis of forms discussed, 326a, 
331c,3,7b,d, 346c,7bc,9d, 351d,2b,4d,5bo, 
7bc, 361cd, 366-7, 376cd, 383a, 386-8, 
390c,5a, 400c, 415d, 420, 422, 423d-5a, 
426bc, 429a,b, 436cd, 447cd, 456, 457cd, 
463bc, 483, 533abc, 538d-9b, 559b-60a. 

Haplographia, 345b,8a, 353c, 563ab. 

% compounds of roots in, 367bc, 453d, 
502a ; final radical t becomes iy, 369c, 
cf. 402-3, 393d. 

Lengthening of final vowels, 332-4, 
338c, 339d, 406d, 411d, 423b,6d, 530b, 
531ab, cf. 538-9, 540a ; of vowels of final 
syllables, 330d, 543b. Cf. Shortening. 

Longer forms to be read in place of 
grammatically equivalent shorter ones, 
345bc, 348ab,c, 357d, 380c, 381c, 393d, 
4d,5d, 403d, 413b, 420c, 427a,9b, 431b, 
458d, 476d, 506c, 514d, 536a, 539d, (cf. 
554a,) 565d; cf. Shorter etc. 

Metre, determines choice between 
grammatically equivalent forms, 535bcd, 
340c, 345a,7cd,9d, 357cd, 548ab. 

Middle forms, their extreme rarity, 
511c; cf. 512c, 514a. 

min, stems in, 542-5. 

Modernization, of accent, 45 5d; of 
form, 342b,5ab,7d,9d, 376d 380a, 388b, 
c,d,9a, 391d, 408a, 430c, 435a, 444b, 
482c, 509b, 513b, 520b, 533b,7d, 554a, 
565a. 

Nasalization, of final vowel, 328a,9a, 
331d,5c,7a,c,8b,9c, 342a, 355d, 470b; of 
stem, see Variation. 

r, compounds of roots in, 502a, 453d ; 
represents a long vowel, 430b. 

Shortening of final vowels, 342bcd,8d, 
355d, 390d,lc,2a, 413b; cf. 413c, 451c, 
638d-40a; cf. Lengthening. 

Shorter forms to be read in place 
of grammatically equivalent longer, 
371d-2a, 383bc, 408a, 420c, 426d, 515a 
(sahyase) ; cf. 345b ; cf . Longer etc. 



Strong and weak cases distinguished, 
see Variation. 

Strong forms for weak (or forms with 
apparently irregular strengthening), 420 
cd, 498d, 516b, 521cd, 523d (cf. 532a,7c, 
8d), 540c, 546c,d ; cf. 428c. 

Supplementary themes, 523ab, 530c- 
31a, 328b, 431ab,2bc, 460c, 462a,6c, 
47lc, 486b,9b,d, 493bc,4c, 516cd, 517b, 
534c: cf. 526-7. 

Syncopation, 523d-5d ; see also Varia- 
tion. 

t, for k, 466a, 456d, 463c, 523b ; for 
t, 499b ; for s, 468b ; inserted between 
-n and s-, 346cd. 

*, for j, 463ab ; for h, 499b ; for p, 
490c ; for sft, 493ab. 

tdt, stems in, 466d. 

Transitions, from consonant to con- 
sonant-declension, 516cd; from conso- 
nant to vowel-declension, 347c, 433d, 
457cd,8d, 462cd, 470a, 471,6a,d,8d,9, 481 
cd,3cd,5c, 486-7,9d, 490a, 491d, 493c-4c, 
8d,9c, 501b,2b,5d, 510a,d,ld,4b,6c, 523a, 
5d-7b, 530a,lft, 542bc,6d,7c, 548d-555b 
(cf. 559ab, 560b,lc,2a,d,3d,4b,c,5a,b,6a, 
c,d,7b,c), 568b-570c (cf. 57lb-573d), 
572d-3a ; from vowel to consonant- 
declension, 347bc, 359b, 373-4, 404, 466 
d, 471c, 484c,6d,494c, 527b, 540d,la,7c, 
8d, 555b-558c (cf. 559ab, 560b,le,2a,d, 
3d,4b,c,5a,b,6a,c.7b), 570c-571a (cf. 571 
c-573d); from vowel to vowel-declen- 
sion. 369d-373 (cf. 375-400 passim), 
403-404 (cf. 405-419 passim), 427a, 
434a, 435d-6d, 439d-40a. 

u, compounds of roots in, 502a, 401cd, 
453d ; final -u in locative regularly 
uncombined, 354, 364-5 ; final radical 
4 becomes uv, 402-3 ; cf. i 

Variation of stem, by added n in 
forms of vowel stems, 331c, 367ab, 373 
cd-4a, 401bc, 404bc ; by n in G.p., 
351d, 364b, 397c, 417d, 430, 432a, 453b; 
by y, 344a, 357c-360b, 361b; by nasali- 
zation, 454c,9c, 462a, 489d, 490b, 504-5, 
516, 545d, 566ab,8a, 573bc; by synco- 
pation, 420c, 478c, 523d-525d', 535ab, 
537d; by vowel-change, 367a, 382-9, 
419-420, 459d, 462b. 470d,2bc, 481b,4c, 
5cd,6b, 493a,8bc. 523bcd, 545d-6d, 566 
ab.8a, 573bc; further. 498b, 511c, 514a. 

pari, feminines in, 527c. 

vin, stems in, 542-5. 

Weak forms for strong (or forms with 
apparently irregular weakening), 420d, 
498c(cf.499d), 509d, 511d,2d,3cd,4b,6ab, 
520d,la,b,3bc,d (cf. 531d,7c,d,8c,d,9c). 



Synopsis.] 



Noun- Inflection in the Veda. 



601 



Synopsis op the fobegodto Article. 



Introduction,. 325 

Stems in derivative a, 329 

In derivative d, 355 

In derivative and radical i and {,._ 365 
In derivative and radical u and &,. 400 

Inror or, _ _ 419 

In ai, o, and au, 431 

In radical A and a, 434 

In radical i, i, u, H, and r, 453 

In radical gh, . ... 454 

In radical ac, 454 

In radical c, 459 

Inradicalcft, 461 

In radical/, 461 

In radical t and d, 466 

In radical *,_. 466 

In radical ih, 470 

In radical d, 410 

In radical dh, 475 



In radical m, 478 

In radical p, 481 

In radical bh, 484 

In radical m, .485 

In radical r, 485 

In radicals, 489 

Inradicalp, 489 

In radical sh and s, 492 

In radical ft, 497 

In final root- vowel plus t, 501 

In nt, t (participles), 504 

In vans (participles), 511 

In yans (comparatives), 514 

In vamt and mant, 515 

In van, man, and an, 522 

In in, 542 

In as, 545 

Tn is and us, 568 



Additional Notes : 1. Dual Forms in -d and -au from Stems not ending in a, 
574 ; 2. The relative Frequency of ancient and modern equivalent grammatical 
Forms as a Criterion of the Age of different "Vedic Texts, 576 ; 3. Numerical 
Summary of the Noun-forms in the Riksanhita^ 581. 

Corrections and Additions, 585 ; Index of Citations, 586 ; Sanskrit Index, 591 ; 
General Index, 599.